diff --git a/evals/team-tasks/vo-t01-llm-judge/environment/repo.tar.gz b/evals/team-tasks/vo-t01-llm-judge/environment/repo.tar.gz index e1ca88d..88e9207 100644 Binary files a/evals/team-tasks/vo-t01-llm-judge/environment/repo.tar.gz and b/evals/team-tasks/vo-t01-llm-judge/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t01-llm-judge/instruction.md b/evals/team-tasks/vo-t01-llm-judge/instruction.md index 5f5e7cf..d82f434 100644 --- a/evals/team-tasks/vo-t01-llm-judge/instruction.md +++ b/evals/team-tasks/vo-t01-llm-judge/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `aa9931fa0eb907c9b622e3faed1a41bed241aa3f` (base) and -`e6c05ff08b755f3e74c9065d036dede7fb5f2158` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `f280a86d34b0121cfd84e0815d6fdafa61f10337` (base) and +`c06bf86bd436dda60bee4c2b69f88f21b39f1ba6` (head). What the change is meant to do: The change adds a Models page for managing local speech models and lets users choose which transcription model the voice pipeline uses. The sidecar gains a registry of named models with categories and recommendations, per-model download status, a delete endpoint, and a transcription model selectable per request. The settings and create-voice screens check only the models needed for the current local mode, and a new config field stores the chosen transcription model. diff --git a/evals/team-tasks/vo-t01-llm-judge/steps/architecture/instruction.md b/evals/team-tasks/vo-t01-llm-judge/steps/architecture/instruction.md index f897df4..e5d938a 100644 --- a/evals/team-tasks/vo-t01-llm-judge/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-t01-llm-judge/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01-llm-judge/steps/code_quality/instruction.md b/evals/team-tasks/vo-t01-llm-judge/steps/code_quality/instruction.md index 1e61894..7e3a23d 100644 --- a/evals/team-tasks/vo-t01-llm-judge/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-t01-llm-judge/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01-llm-judge/steps/compliance/instruction.md b/evals/team-tasks/vo-t01-llm-judge/steps/compliance/instruction.md index 9981db7..2d5c5f2 100644 --- a/evals/team-tasks/vo-t01-llm-judge/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-t01-llm-judge/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01-llm-judge/steps/correctness/instruction.md b/evals/team-tasks/vo-t01-llm-judge/steps/correctness/instruction.md index 5f4ad2c..ce01551 100644 --- a/evals/team-tasks/vo-t01-llm-judge/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-t01-llm-judge/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01-llm-judge/steps/language/instruction.md b/evals/team-tasks/vo-t01-llm-judge/steps/language/instruction.md index e03279a..cb87d5f 100644 --- a/evals/team-tasks/vo-t01-llm-judge/steps/language/instruction.md +++ b/evals/team-tasks/vo-t01-llm-judge/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01-llm-judge/steps/security/instruction.md b/evals/team-tasks/vo-t01-llm-judge/steps/security/instruction.md index 6f415ab..e80530a 100644 --- a/evals/team-tasks/vo-t01-llm-judge/steps/security/instruction.md +++ b/evals/team-tasks/vo-t01-llm-judge/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01-llm-judge/steps/tests/instruction.md b/evals/team-tasks/vo-t01-llm-judge/steps/tests/instruction.md index 56d1944..bfe9375 100644 --- a/evals/team-tasks/vo-t01-llm-judge/steps/tests/instruction.md +++ b/evals/team-tasks/vo-t01-llm-judge/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01-llm-judge/steps/validator/instruction.md b/evals/team-tasks/vo-t01-llm-judge/steps/validator/instruction.md index 879c33e..7fe2cf5 100644 --- a/evals/team-tasks/vo-t01-llm-judge/steps/validator/instruction.md +++ b/evals/team-tasks/vo-t01-llm-judge/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is aa9931fa0eb907c9b622e3faed1a41bed241aa3f. +Use the review-validator skill. The output folder is /workspace/review. The base ref is f280a86d34b0121cfd84e0815d6fdafa61f10337. diff --git a/evals/team-tasks/vo-t01-llm-judge/steps/validator/tests/workspace/repo.tar.gz b/evals/team-tasks/vo-t01-llm-judge/steps/validator/tests/workspace/repo.tar.gz index e1ca88d..88e9207 100644 Binary files a/evals/team-tasks/vo-t01-llm-judge/steps/validator/tests/workspace/repo.tar.gz and b/evals/team-tasks/vo-t01-llm-judge/steps/validator/tests/workspace/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t01-llm-judge/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-t01-llm-judge/steps/validator/workdir/setup.sh index 09d1387..6e06787 100755 --- a/evals/team-tasks/vo-t01-llm-judge/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-t01-llm-judge/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base aa9931fa0eb907c9b622e3faed1a41bed241aa3f --head e6c05ff08b755f3e74c9065d036dede7fb5f2158 +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base f280a86d34b0121cfd84e0815d6fdafa61f10337 --head c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 diff --git a/evals/team-tasks/vo-t01/environment/repo.tar.gz b/evals/team-tasks/vo-t01/environment/repo.tar.gz index e1ca88d..88e9207 100644 Binary files a/evals/team-tasks/vo-t01/environment/repo.tar.gz and b/evals/team-tasks/vo-t01/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t01/instruction.md b/evals/team-tasks/vo-t01/instruction.md index 5f5e7cf..d82f434 100644 --- a/evals/team-tasks/vo-t01/instruction.md +++ b/evals/team-tasks/vo-t01/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `aa9931fa0eb907c9b622e3faed1a41bed241aa3f` (base) and -`e6c05ff08b755f3e74c9065d036dede7fb5f2158` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `f280a86d34b0121cfd84e0815d6fdafa61f10337` (base) and +`c06bf86bd436dda60bee4c2b69f88f21b39f1ba6` (head). What the change is meant to do: The change adds a Models page for managing local speech models and lets users choose which transcription model the voice pipeline uses. The sidecar gains a registry of named models with categories and recommendations, per-model download status, a delete endpoint, and a transcription model selectable per request. The settings and create-voice screens check only the models needed for the current local mode, and a new config field stores the chosen transcription model. diff --git a/evals/team-tasks/vo-t01/steps/architecture/instruction.md b/evals/team-tasks/vo-t01/steps/architecture/instruction.md index f897df4..e5d938a 100644 --- a/evals/team-tasks/vo-t01/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-t01/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01/steps/code_quality/instruction.md b/evals/team-tasks/vo-t01/steps/code_quality/instruction.md index 1e61894..7e3a23d 100644 --- a/evals/team-tasks/vo-t01/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-t01/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01/steps/compliance/instruction.md b/evals/team-tasks/vo-t01/steps/compliance/instruction.md index 9981db7..2d5c5f2 100644 --- a/evals/team-tasks/vo-t01/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-t01/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01/steps/correctness/instruction.md b/evals/team-tasks/vo-t01/steps/correctness/instruction.md index 5f4ad2c..ce01551 100644 --- a/evals/team-tasks/vo-t01/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-t01/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01/steps/language/instruction.md b/evals/team-tasks/vo-t01/steps/language/instruction.md index e03279a..cb87d5f 100644 --- a/evals/team-tasks/vo-t01/steps/language/instruction.md +++ b/evals/team-tasks/vo-t01/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01/steps/security/instruction.md b/evals/team-tasks/vo-t01/steps/security/instruction.md index 6f415ab..e80530a 100644 --- a/evals/team-tasks/vo-t01/steps/security/instruction.md +++ b/evals/team-tasks/vo-t01/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01/steps/tests/instruction.md b/evals/team-tasks/vo-t01/steps/tests/instruction.md index 56d1944..bfe9375 100644 --- a/evals/team-tasks/vo-t01/steps/tests/instruction.md +++ b/evals/team-tasks/vo-t01/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs aa9931fa0eb907c9b622e3faed1a41bed241aa3f and e6c05ff08b755f3e74c9065d036dede7fb5f2158 in the repository in /workspace/repo. +Review the change between git refs f280a86d34b0121cfd84e0815d6fdafa61f10337 and c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t01/steps/validator/instruction.md b/evals/team-tasks/vo-t01/steps/validator/instruction.md index 879c33e..7fe2cf5 100644 --- a/evals/team-tasks/vo-t01/steps/validator/instruction.md +++ b/evals/team-tasks/vo-t01/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is aa9931fa0eb907c9b622e3faed1a41bed241aa3f. +Use the review-validator skill. The output folder is /workspace/review. The base ref is f280a86d34b0121cfd84e0815d6fdafa61f10337. diff --git a/evals/team-tasks/vo-t01/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-t01/steps/validator/workdir/setup.sh index 09d1387..6e06787 100755 --- a/evals/team-tasks/vo-t01/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-t01/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base aa9931fa0eb907c9b622e3faed1a41bed241aa3f --head e6c05ff08b755f3e74c9065d036dede7fb5f2158 +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base f280a86d34b0121cfd84e0815d6fdafa61f10337 --head c06bf86bd436dda60bee4c2b69f88f21b39f1ba6 diff --git a/evals/team-tasks/vo-t02-llm-judge/environment/repo.tar.gz b/evals/team-tasks/vo-t02-llm-judge/environment/repo.tar.gz index c6f1b85..e3b47d2 100644 Binary files a/evals/team-tasks/vo-t02-llm-judge/environment/repo.tar.gz and b/evals/team-tasks/vo-t02-llm-judge/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t02-llm-judge/instruction.md b/evals/team-tasks/vo-t02-llm-judge/instruction.md index 2868388..ec90211 100644 --- a/evals/team-tasks/vo-t02-llm-judge/instruction.md +++ b/evals/team-tasks/vo-t02-llm-judge/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `95142b184077306b75e5957c0462c00e40a86e0b` (base) and -`35892f2d11ed26e6830b20cae8125cbc25287a9d` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1` (base) and +`9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1` (head). What the change is meant to do: The change adds a bundled local text-to-speech sidecar that replaces the dependency on an external local voice server. A Python service handles transcription, voice profiles, chunked speech generation, model downloads and YouTube audio extraction, while the Rust backend starts, health-checks and stops it on a managed port and exposes commands to reach it. The frontend client, home, settings and create-voice screens route through these commands, and the build and CI setup include the sidecar binary. diff --git a/evals/team-tasks/vo-t02-llm-judge/steps/architecture/instruction.md b/evals/team-tasks/vo-t02-llm-judge/steps/architecture/instruction.md index f25bd26..9d495bb 100644 --- a/evals/team-tasks/vo-t02-llm-judge/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-t02-llm-judge/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02-llm-judge/steps/code_quality/instruction.md b/evals/team-tasks/vo-t02-llm-judge/steps/code_quality/instruction.md index 41bec19..d435e5b 100644 --- a/evals/team-tasks/vo-t02-llm-judge/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-t02-llm-judge/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02-llm-judge/steps/compliance/instruction.md b/evals/team-tasks/vo-t02-llm-judge/steps/compliance/instruction.md index 1128fd6..4ed1e7c 100644 --- a/evals/team-tasks/vo-t02-llm-judge/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-t02-llm-judge/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02-llm-judge/steps/correctness/instruction.md b/evals/team-tasks/vo-t02-llm-judge/steps/correctness/instruction.md index b48df41..fd061c1 100644 --- a/evals/team-tasks/vo-t02-llm-judge/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-t02-llm-judge/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02-llm-judge/steps/language/instruction.md b/evals/team-tasks/vo-t02-llm-judge/steps/language/instruction.md index 9557352..1dbd550 100644 --- a/evals/team-tasks/vo-t02-llm-judge/steps/language/instruction.md +++ b/evals/team-tasks/vo-t02-llm-judge/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02-llm-judge/steps/security/instruction.md b/evals/team-tasks/vo-t02-llm-judge/steps/security/instruction.md index 15877fc..4eb0532 100644 --- a/evals/team-tasks/vo-t02-llm-judge/steps/security/instruction.md +++ b/evals/team-tasks/vo-t02-llm-judge/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02-llm-judge/steps/tests/instruction.md b/evals/team-tasks/vo-t02-llm-judge/steps/tests/instruction.md index 26b7954..5a11b89 100644 --- a/evals/team-tasks/vo-t02-llm-judge/steps/tests/instruction.md +++ b/evals/team-tasks/vo-t02-llm-judge/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02-llm-judge/steps/validator/instruction.md b/evals/team-tasks/vo-t02-llm-judge/steps/validator/instruction.md index 25e0c26..31e87d2 100644 --- a/evals/team-tasks/vo-t02-llm-judge/steps/validator/instruction.md +++ b/evals/team-tasks/vo-t02-llm-judge/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is 95142b184077306b75e5957c0462c00e40a86e0b. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1. diff --git a/evals/team-tasks/vo-t02-llm-judge/steps/validator/tests/workspace/repo.tar.gz b/evals/team-tasks/vo-t02-llm-judge/steps/validator/tests/workspace/repo.tar.gz index c6f1b85..e3b47d2 100644 Binary files a/evals/team-tasks/vo-t02-llm-judge/steps/validator/tests/workspace/repo.tar.gz and b/evals/team-tasks/vo-t02-llm-judge/steps/validator/tests/workspace/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t02-llm-judge/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-t02-llm-judge/steps/validator/workdir/setup.sh index 5952eb8..9502489 100755 --- a/evals/team-tasks/vo-t02-llm-judge/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-t02-llm-judge/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 95142b184077306b75e5957c0462c00e40a86e0b --head 35892f2d11ed26e6830b20cae8125cbc25287a9d +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 --head 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 diff --git a/evals/team-tasks/vo-t02/environment/repo.tar.gz b/evals/team-tasks/vo-t02/environment/repo.tar.gz index c6f1b85..e3b47d2 100644 Binary files a/evals/team-tasks/vo-t02/environment/repo.tar.gz and b/evals/team-tasks/vo-t02/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t02/instruction.md b/evals/team-tasks/vo-t02/instruction.md index 2868388..ec90211 100644 --- a/evals/team-tasks/vo-t02/instruction.md +++ b/evals/team-tasks/vo-t02/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `95142b184077306b75e5957c0462c00e40a86e0b` (base) and -`35892f2d11ed26e6830b20cae8125cbc25287a9d` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1` (base) and +`9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1` (head). What the change is meant to do: The change adds a bundled local text-to-speech sidecar that replaces the dependency on an external local voice server. A Python service handles transcription, voice profiles, chunked speech generation, model downloads and YouTube audio extraction, while the Rust backend starts, health-checks and stops it on a managed port and exposes commands to reach it. The frontend client, home, settings and create-voice screens route through these commands, and the build and CI setup include the sidecar binary. diff --git a/evals/team-tasks/vo-t02/steps/architecture/instruction.md b/evals/team-tasks/vo-t02/steps/architecture/instruction.md index f25bd26..9d495bb 100644 --- a/evals/team-tasks/vo-t02/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-t02/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02/steps/code_quality/instruction.md b/evals/team-tasks/vo-t02/steps/code_quality/instruction.md index 41bec19..d435e5b 100644 --- a/evals/team-tasks/vo-t02/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-t02/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02/steps/compliance/instruction.md b/evals/team-tasks/vo-t02/steps/compliance/instruction.md index 1128fd6..4ed1e7c 100644 --- a/evals/team-tasks/vo-t02/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-t02/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02/steps/correctness/instruction.md b/evals/team-tasks/vo-t02/steps/correctness/instruction.md index b48df41..fd061c1 100644 --- a/evals/team-tasks/vo-t02/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-t02/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02/steps/language/instruction.md b/evals/team-tasks/vo-t02/steps/language/instruction.md index 9557352..1dbd550 100644 --- a/evals/team-tasks/vo-t02/steps/language/instruction.md +++ b/evals/team-tasks/vo-t02/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02/steps/security/instruction.md b/evals/team-tasks/vo-t02/steps/security/instruction.md index 15877fc..4eb0532 100644 --- a/evals/team-tasks/vo-t02/steps/security/instruction.md +++ b/evals/team-tasks/vo-t02/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02/steps/tests/instruction.md b/evals/team-tasks/vo-t02/steps/tests/instruction.md index 26b7954..5a11b89 100644 --- a/evals/team-tasks/vo-t02/steps/tests/instruction.md +++ b/evals/team-tasks/vo-t02/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs 95142b184077306b75e5957c0462c00e40a86e0b and 35892f2d11ed26e6830b20cae8125cbc25287a9d in the repository in /workspace/repo. +Review the change between git refs 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 and 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t02/steps/validator/instruction.md b/evals/team-tasks/vo-t02/steps/validator/instruction.md index 25e0c26..31e87d2 100644 --- a/evals/team-tasks/vo-t02/steps/validator/instruction.md +++ b/evals/team-tasks/vo-t02/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is 95142b184077306b75e5957c0462c00e40a86e0b. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1. diff --git a/evals/team-tasks/vo-t02/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-t02/steps/validator/workdir/setup.sh index 5952eb8..9502489 100755 --- a/evals/team-tasks/vo-t02/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-t02/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 95142b184077306b75e5957c0462c00e40a86e0b --head 35892f2d11ed26e6830b20cae8125cbc25287a9d +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 3f2a14268a65847b7cc5a8961eb7fb8a4bf575b1 --head 9afeb65d8f0b0262250a7e8a11f146f4bf0e1eb1 diff --git a/evals/team-tasks/vo-t03/environment/repo.tar.gz b/evals/team-tasks/vo-t03/environment/repo.tar.gz index 453beff..0d6dba7 100644 Binary files a/evals/team-tasks/vo-t03/environment/repo.tar.gz and b/evals/team-tasks/vo-t03/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t03/instruction.md b/evals/team-tasks/vo-t03/instruction.md index 744ac21..1e7bef4 100644 --- a/evals/team-tasks/vo-t03/instruction.md +++ b/evals/team-tasks/vo-t03/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `929123b134d41df41a57942ab81376335e18ccbd` (base) and -`34f3aef77486fd1c6666bcb9580ba45b5cd0c128` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `0b3d1346cb747cee4647c144d91c955f9522cf34` (base) and +`3a944f6c5bd372b00769eeecd950eb180cbc656b` (head). What the change is meant to do: The change extends the local voice sidecar with a voice-conversion mode alongside text-to-speech, and makes generated audio follow the original recording's timing. The sidecar gains a conversion endpoint, timestamp-aligned segment generation, transcript filtering and path validation; the Rust backend probes video duration, picks the mode from config, saves a transcript next to the output and cleans up stale temporary recordings. It also adds sidecar test suites, a build script and an expanded README. diff --git a/evals/team-tasks/vo-t03/steps/architecture/instruction.md b/evals/team-tasks/vo-t03/steps/architecture/instruction.md index 4493c07..c6533bc 100644 --- a/evals/team-tasks/vo-t03/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-t03/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs 929123b134d41df41a57942ab81376335e18ccbd and 34f3aef77486fd1c6666bcb9580ba45b5cd0c128 in the repository in /workspace/repo. +Review the change between git refs 0b3d1346cb747cee4647c144d91c955f9522cf34 and 3a944f6c5bd372b00769eeecd950eb180cbc656b in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t03/steps/code_quality/instruction.md b/evals/team-tasks/vo-t03/steps/code_quality/instruction.md index 77fb99e..0e4ff63 100644 --- a/evals/team-tasks/vo-t03/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-t03/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs 929123b134d41df41a57942ab81376335e18ccbd and 34f3aef77486fd1c6666bcb9580ba45b5cd0c128 in the repository in /workspace/repo. +Review the change between git refs 0b3d1346cb747cee4647c144d91c955f9522cf34 and 3a944f6c5bd372b00769eeecd950eb180cbc656b in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t03/steps/compliance/instruction.md b/evals/team-tasks/vo-t03/steps/compliance/instruction.md index ca9d183..55c8813 100644 --- a/evals/team-tasks/vo-t03/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-t03/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs 929123b134d41df41a57942ab81376335e18ccbd and 34f3aef77486fd1c6666bcb9580ba45b5cd0c128 in the repository in /workspace/repo. +Review the change between git refs 0b3d1346cb747cee4647c144d91c955f9522cf34 and 3a944f6c5bd372b00769eeecd950eb180cbc656b in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t03/steps/correctness/instruction.md b/evals/team-tasks/vo-t03/steps/correctness/instruction.md index eaf78f9..62edd90 100644 --- a/evals/team-tasks/vo-t03/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-t03/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs 929123b134d41df41a57942ab81376335e18ccbd and 34f3aef77486fd1c6666bcb9580ba45b5cd0c128 in the repository in /workspace/repo. +Review the change between git refs 0b3d1346cb747cee4647c144d91c955f9522cf34 and 3a944f6c5bd372b00769eeecd950eb180cbc656b in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t03/steps/language/instruction.md b/evals/team-tasks/vo-t03/steps/language/instruction.md index e854eeb..5459e55 100644 --- a/evals/team-tasks/vo-t03/steps/language/instruction.md +++ b/evals/team-tasks/vo-t03/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs 929123b134d41df41a57942ab81376335e18ccbd and 34f3aef77486fd1c6666bcb9580ba45b5cd0c128 in the repository in /workspace/repo. +Review the change between git refs 0b3d1346cb747cee4647c144d91c955f9522cf34 and 3a944f6c5bd372b00769eeecd950eb180cbc656b in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t03/steps/security/instruction.md b/evals/team-tasks/vo-t03/steps/security/instruction.md index 425fbd9..004e0e9 100644 --- a/evals/team-tasks/vo-t03/steps/security/instruction.md +++ b/evals/team-tasks/vo-t03/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs 929123b134d41df41a57942ab81376335e18ccbd and 34f3aef77486fd1c6666bcb9580ba45b5cd0c128 in the repository in /workspace/repo. +Review the change between git refs 0b3d1346cb747cee4647c144d91c955f9522cf34 and 3a944f6c5bd372b00769eeecd950eb180cbc656b in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t03/steps/tests/instruction.md b/evals/team-tasks/vo-t03/steps/tests/instruction.md index 9f84546..7bd1269 100644 --- a/evals/team-tasks/vo-t03/steps/tests/instruction.md +++ b/evals/team-tasks/vo-t03/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs 929123b134d41df41a57942ab81376335e18ccbd and 34f3aef77486fd1c6666bcb9580ba45b5cd0c128 in the repository in /workspace/repo. +Review the change between git refs 0b3d1346cb747cee4647c144d91c955f9522cf34 and 3a944f6c5bd372b00769eeecd950eb180cbc656b in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t03/steps/validator/instruction.md b/evals/team-tasks/vo-t03/steps/validator/instruction.md index 99267e0..8580b46 100644 --- a/evals/team-tasks/vo-t03/steps/validator/instruction.md +++ b/evals/team-tasks/vo-t03/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is 929123b134d41df41a57942ab81376335e18ccbd. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 0b3d1346cb747cee4647c144d91c955f9522cf34. diff --git a/evals/team-tasks/vo-t03/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-t03/steps/validator/workdir/setup.sh index c10ed37..9030e4d 100755 --- a/evals/team-tasks/vo-t03/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-t03/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 929123b134d41df41a57942ab81376335e18ccbd --head 34f3aef77486fd1c6666bcb9580ba45b5cd0c128 +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 0b3d1346cb747cee4647c144d91c955f9522cf34 --head 3a944f6c5bd372b00769eeecd950eb180cbc656b diff --git a/evals/team-tasks/vo-t04-llm-judge/environment/repo.tar.gz b/evals/team-tasks/vo-t04-llm-judge/environment/repo.tar.gz index 5b21220..30a2d66 100644 Binary files a/evals/team-tasks/vo-t04-llm-judge/environment/repo.tar.gz and b/evals/team-tasks/vo-t04-llm-judge/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t04-llm-judge/instruction.md b/evals/team-tasks/vo-t04-llm-judge/instruction.md index 9298cdf..76ac89d 100644 --- a/evals/team-tasks/vo-t04-llm-judge/instruction.md +++ b/evals/team-tasks/vo-t04-llm-judge/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `099656be3d2d779b7908a35e468feb7fe98b0619` (base) and -`b9d28e98158fc7b10366fee4286252ff65a9dc0b` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `15fc88055ae2fb059db8e5b001b5807a14fdbcbd` (base) and +`152365b8573f1f24309de7ab6366a5c75b56ed81` (head). What the change is meant to do: The change adds a webcam overlay to recordings and moves stored credentials into the operating system keychain. The recorder composites the webcam bubble into the recorded video, with a preview bubble that can switch between two corner positions saved in preferences. Config handling saves secrets to the keychain, removes them from the config file and local storage, and limits file reads to allowed folders; the Drive sign-in now carries a state value. It also adds a CI workflow, an icon generator script and an in-memory blob store. diff --git a/evals/team-tasks/vo-t04-llm-judge/steps/architecture/instruction.md b/evals/team-tasks/vo-t04-llm-judge/steps/architecture/instruction.md index 18766e7..e9279ff 100644 --- a/evals/team-tasks/vo-t04-llm-judge/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-t04-llm-judge/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04-llm-judge/steps/code_quality/instruction.md b/evals/team-tasks/vo-t04-llm-judge/steps/code_quality/instruction.md index 5c7fff7..43113b0 100644 --- a/evals/team-tasks/vo-t04-llm-judge/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-t04-llm-judge/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04-llm-judge/steps/compliance/instruction.md b/evals/team-tasks/vo-t04-llm-judge/steps/compliance/instruction.md index f30727e..8faee6e 100644 --- a/evals/team-tasks/vo-t04-llm-judge/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-t04-llm-judge/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04-llm-judge/steps/correctness/instruction.md b/evals/team-tasks/vo-t04-llm-judge/steps/correctness/instruction.md index be84a7f..b11598a 100644 --- a/evals/team-tasks/vo-t04-llm-judge/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-t04-llm-judge/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04-llm-judge/steps/language/instruction.md b/evals/team-tasks/vo-t04-llm-judge/steps/language/instruction.md index d39960c..c550cfb 100644 --- a/evals/team-tasks/vo-t04-llm-judge/steps/language/instruction.md +++ b/evals/team-tasks/vo-t04-llm-judge/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04-llm-judge/steps/security/instruction.md b/evals/team-tasks/vo-t04-llm-judge/steps/security/instruction.md index 78a6954..b7a4885 100644 --- a/evals/team-tasks/vo-t04-llm-judge/steps/security/instruction.md +++ b/evals/team-tasks/vo-t04-llm-judge/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04-llm-judge/steps/tests/instruction.md b/evals/team-tasks/vo-t04-llm-judge/steps/tests/instruction.md index 61af110..3ed0d3b 100644 --- a/evals/team-tasks/vo-t04-llm-judge/steps/tests/instruction.md +++ b/evals/team-tasks/vo-t04-llm-judge/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04-llm-judge/steps/validator/instruction.md b/evals/team-tasks/vo-t04-llm-judge/steps/validator/instruction.md index 0757776..13e490d 100644 --- a/evals/team-tasks/vo-t04-llm-judge/steps/validator/instruction.md +++ b/evals/team-tasks/vo-t04-llm-judge/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is 099656be3d2d779b7908a35e468feb7fe98b0619. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 15fc88055ae2fb059db8e5b001b5807a14fdbcbd. diff --git a/evals/team-tasks/vo-t04-llm-judge/steps/validator/tests/workspace/repo.tar.gz b/evals/team-tasks/vo-t04-llm-judge/steps/validator/tests/workspace/repo.tar.gz index 5b21220..30a2d66 100644 Binary files a/evals/team-tasks/vo-t04-llm-judge/steps/validator/tests/workspace/repo.tar.gz and b/evals/team-tasks/vo-t04-llm-judge/steps/validator/tests/workspace/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t04-llm-judge/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-t04-llm-judge/steps/validator/workdir/setup.sh index d5f8578..6eb04f0 100755 --- a/evals/team-tasks/vo-t04-llm-judge/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-t04-llm-judge/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 099656be3d2d779b7908a35e468feb7fe98b0619 --head b9d28e98158fc7b10366fee4286252ff65a9dc0b +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 15fc88055ae2fb059db8e5b001b5807a14fdbcbd --head 152365b8573f1f24309de7ab6366a5c75b56ed81 diff --git a/evals/team-tasks/vo-t04/environment/repo.tar.gz b/evals/team-tasks/vo-t04/environment/repo.tar.gz index 5b21220..30a2d66 100644 Binary files a/evals/team-tasks/vo-t04/environment/repo.tar.gz and b/evals/team-tasks/vo-t04/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t04/instruction.md b/evals/team-tasks/vo-t04/instruction.md index 9298cdf..76ac89d 100644 --- a/evals/team-tasks/vo-t04/instruction.md +++ b/evals/team-tasks/vo-t04/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `099656be3d2d779b7908a35e468feb7fe98b0619` (base) and -`b9d28e98158fc7b10366fee4286252ff65a9dc0b` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `15fc88055ae2fb059db8e5b001b5807a14fdbcbd` (base) and +`152365b8573f1f24309de7ab6366a5c75b56ed81` (head). What the change is meant to do: The change adds a webcam overlay to recordings and moves stored credentials into the operating system keychain. The recorder composites the webcam bubble into the recorded video, with a preview bubble that can switch between two corner positions saved in preferences. Config handling saves secrets to the keychain, removes them from the config file and local storage, and limits file reads to allowed folders; the Drive sign-in now carries a state value. It also adds a CI workflow, an icon generator script and an in-memory blob store. diff --git a/evals/team-tasks/vo-t04/steps/architecture/instruction.md b/evals/team-tasks/vo-t04/steps/architecture/instruction.md index 18766e7..e9279ff 100644 --- a/evals/team-tasks/vo-t04/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-t04/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04/steps/code_quality/instruction.md b/evals/team-tasks/vo-t04/steps/code_quality/instruction.md index 5c7fff7..43113b0 100644 --- a/evals/team-tasks/vo-t04/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-t04/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04/steps/compliance/instruction.md b/evals/team-tasks/vo-t04/steps/compliance/instruction.md index f30727e..8faee6e 100644 --- a/evals/team-tasks/vo-t04/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-t04/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04/steps/correctness/instruction.md b/evals/team-tasks/vo-t04/steps/correctness/instruction.md index be84a7f..b11598a 100644 --- a/evals/team-tasks/vo-t04/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-t04/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04/steps/language/instruction.md b/evals/team-tasks/vo-t04/steps/language/instruction.md index d39960c..c550cfb 100644 --- a/evals/team-tasks/vo-t04/steps/language/instruction.md +++ b/evals/team-tasks/vo-t04/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04/steps/security/instruction.md b/evals/team-tasks/vo-t04/steps/security/instruction.md index 78a6954..b7a4885 100644 --- a/evals/team-tasks/vo-t04/steps/security/instruction.md +++ b/evals/team-tasks/vo-t04/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04/steps/tests/instruction.md b/evals/team-tasks/vo-t04/steps/tests/instruction.md index 61af110..3ed0d3b 100644 --- a/evals/team-tasks/vo-t04/steps/tests/instruction.md +++ b/evals/team-tasks/vo-t04/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs 099656be3d2d779b7908a35e468feb7fe98b0619 and b9d28e98158fc7b10366fee4286252ff65a9dc0b in the repository in /workspace/repo. +Review the change between git refs 15fc88055ae2fb059db8e5b001b5807a14fdbcbd and 152365b8573f1f24309de7ab6366a5c75b56ed81 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t04/steps/validator/instruction.md b/evals/team-tasks/vo-t04/steps/validator/instruction.md index 0757776..13e490d 100644 --- a/evals/team-tasks/vo-t04/steps/validator/instruction.md +++ b/evals/team-tasks/vo-t04/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is 099656be3d2d779b7908a35e468feb7fe98b0619. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 15fc88055ae2fb059db8e5b001b5807a14fdbcbd. diff --git a/evals/team-tasks/vo-t04/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-t04/steps/validator/workdir/setup.sh index d5f8578..6eb04f0 100755 --- a/evals/team-tasks/vo-t04/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-t04/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 099656be3d2d779b7908a35e468feb7fe98b0619 --head b9d28e98158fc7b10366fee4286252ff65a9dc0b +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 15fc88055ae2fb059db8e5b001b5807a14fdbcbd --head 152365b8573f1f24309de7ab6366a5c75b56ed81 diff --git a/evals/team-tasks/vo-t05/environment/repo.tar.gz b/evals/team-tasks/vo-t05/environment/repo.tar.gz index b5d5f71..2422995 100644 Binary files a/evals/team-tasks/vo-t05/environment/repo.tar.gz and b/evals/team-tasks/vo-t05/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t05/instruction.md b/evals/team-tasks/vo-t05/instruction.md index 17ed3a5..0f27fb4 100644 --- a/evals/team-tasks/vo-t05/instruction.md +++ b/evals/team-tasks/vo-t05/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `ae7315cffaf8c190c80dc6c3288f4d14c0d6ac4e` (base) and -`6e16cfae8882cba5ba1f59edb4d434c74513185a` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `097979928cd0a5376e81f2129cf81d50d425a70d` (base) and +`471de15e63ce9030beee4560c38a9fc45c7d9837` (head). What the change is meant to do: The change adds a Library screen and a sidebar layout for browsing past recordings. The Rust backend lists recordings with sizes, dates, thumbnails and metadata, and offers commands to generate thumbnails, delete, open and reveal files; the pipeline and Drive upload write a companion metadata file with voice and share-link details. Existing pages move under a shared layout with a sidebar and status bar, and the main window is made wider. diff --git a/evals/team-tasks/vo-t05/steps/architecture/instruction.md b/evals/team-tasks/vo-t05/steps/architecture/instruction.md index 459caa9..75de439 100644 --- a/evals/team-tasks/vo-t05/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-t05/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs ae7315cffaf8c190c80dc6c3288f4d14c0d6ac4e and 6e16cfae8882cba5ba1f59edb4d434c74513185a in the repository in /workspace/repo. +Review the change between git refs 097979928cd0a5376e81f2129cf81d50d425a70d and 471de15e63ce9030beee4560c38a9fc45c7d9837 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t05/steps/code_quality/instruction.md b/evals/team-tasks/vo-t05/steps/code_quality/instruction.md index 431a61e..3719ff4 100644 --- a/evals/team-tasks/vo-t05/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-t05/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs ae7315cffaf8c190c80dc6c3288f4d14c0d6ac4e and 6e16cfae8882cba5ba1f59edb4d434c74513185a in the repository in /workspace/repo. +Review the change between git refs 097979928cd0a5376e81f2129cf81d50d425a70d and 471de15e63ce9030beee4560c38a9fc45c7d9837 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t05/steps/compliance/instruction.md b/evals/team-tasks/vo-t05/steps/compliance/instruction.md index 00ffac2..cde0611 100644 --- a/evals/team-tasks/vo-t05/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-t05/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs ae7315cffaf8c190c80dc6c3288f4d14c0d6ac4e and 6e16cfae8882cba5ba1f59edb4d434c74513185a in the repository in /workspace/repo. +Review the change between git refs 097979928cd0a5376e81f2129cf81d50d425a70d and 471de15e63ce9030beee4560c38a9fc45c7d9837 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t05/steps/correctness/instruction.md b/evals/team-tasks/vo-t05/steps/correctness/instruction.md index 3c1ee52..7fc8f4a 100644 --- a/evals/team-tasks/vo-t05/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-t05/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs ae7315cffaf8c190c80dc6c3288f4d14c0d6ac4e and 6e16cfae8882cba5ba1f59edb4d434c74513185a in the repository in /workspace/repo. +Review the change between git refs 097979928cd0a5376e81f2129cf81d50d425a70d and 471de15e63ce9030beee4560c38a9fc45c7d9837 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t05/steps/language/instruction.md b/evals/team-tasks/vo-t05/steps/language/instruction.md index a971836..c0817d1 100644 --- a/evals/team-tasks/vo-t05/steps/language/instruction.md +++ b/evals/team-tasks/vo-t05/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs ae7315cffaf8c190c80dc6c3288f4d14c0d6ac4e and 6e16cfae8882cba5ba1f59edb4d434c74513185a in the repository in /workspace/repo. +Review the change between git refs 097979928cd0a5376e81f2129cf81d50d425a70d and 471de15e63ce9030beee4560c38a9fc45c7d9837 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t05/steps/security/instruction.md b/evals/team-tasks/vo-t05/steps/security/instruction.md index 653d75d..7ee926e 100644 --- a/evals/team-tasks/vo-t05/steps/security/instruction.md +++ b/evals/team-tasks/vo-t05/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs ae7315cffaf8c190c80dc6c3288f4d14c0d6ac4e and 6e16cfae8882cba5ba1f59edb4d434c74513185a in the repository in /workspace/repo. +Review the change between git refs 097979928cd0a5376e81f2129cf81d50d425a70d and 471de15e63ce9030beee4560c38a9fc45c7d9837 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t05/steps/tests/instruction.md b/evals/team-tasks/vo-t05/steps/tests/instruction.md index 534e24e..51f2b57 100644 --- a/evals/team-tasks/vo-t05/steps/tests/instruction.md +++ b/evals/team-tasks/vo-t05/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs ae7315cffaf8c190c80dc6c3288f4d14c0d6ac4e and 6e16cfae8882cba5ba1f59edb4d434c74513185a in the repository in /workspace/repo. +Review the change between git refs 097979928cd0a5376e81f2129cf81d50d425a70d and 471de15e63ce9030beee4560c38a9fc45c7d9837 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t05/steps/validator/instruction.md b/evals/team-tasks/vo-t05/steps/validator/instruction.md index fa8a1f9..f758bfc 100644 --- a/evals/team-tasks/vo-t05/steps/validator/instruction.md +++ b/evals/team-tasks/vo-t05/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is ae7315cffaf8c190c80dc6c3288f4d14c0d6ac4e. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 097979928cd0a5376e81f2129cf81d50d425a70d. diff --git a/evals/team-tasks/vo-t05/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-t05/steps/validator/workdir/setup.sh index 2d92498..5541f7e 100755 --- a/evals/team-tasks/vo-t05/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-t05/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base ae7315cffaf8c190c80dc6c3288f4d14c0d6ac4e --head 6e16cfae8882cba5ba1f59edb4d434c74513185a +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 097979928cd0a5376e81f2129cf81d50d425a70d --head 471de15e63ce9030beee4560c38a9fc45c7d9837 diff --git a/evals/team-tasks/vo-t06-llm-judge/environment/repo.tar.gz b/evals/team-tasks/vo-t06-llm-judge/environment/repo.tar.gz index 32f5d55..45a686d 100644 Binary files a/evals/team-tasks/vo-t06-llm-judge/environment/repo.tar.gz and b/evals/team-tasks/vo-t06-llm-judge/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t06-llm-judge/instruction.md b/evals/team-tasks/vo-t06-llm-judge/instruction.md index c468af4..ea5c375 100644 --- a/evals/team-tasks/vo-t06-llm-judge/instruction.md +++ b/evals/team-tasks/vo-t06-llm-judge/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `12998e404b8d505056221802db4756d39adbd5a0` (base) and -`b27e76080c69c9fe7eb127e1db7f1f9ae692088a` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `d47ca8270bc57385662f40f0c12f83f92b572c96` (base) and +`d9677189b6880ce609fd32f9132b8fc965fac999` (head). What the change is meant to do: The change adds region capture to screen recording. After the user picks a screen, a selection overlay shows a captured frame and lets them draw a rectangle, and the recorder crops the stream to that rectangle. It also maps the capture mode to the picker's surface hint, minimizes and restores the app window around the picker, waits for the first video frame before compositing, and sets a dark background color on the window and page. diff --git a/evals/team-tasks/vo-t06-llm-judge/steps/architecture/instruction.md b/evals/team-tasks/vo-t06-llm-judge/steps/architecture/instruction.md index f0a1014..910ad0b 100644 --- a/evals/team-tasks/vo-t06-llm-judge/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-t06-llm-judge/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06-llm-judge/steps/code_quality/instruction.md b/evals/team-tasks/vo-t06-llm-judge/steps/code_quality/instruction.md index 2dfc5aa..efa1969 100644 --- a/evals/team-tasks/vo-t06-llm-judge/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-t06-llm-judge/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06-llm-judge/steps/compliance/instruction.md b/evals/team-tasks/vo-t06-llm-judge/steps/compliance/instruction.md index 3c0c51f..12aab79 100644 --- a/evals/team-tasks/vo-t06-llm-judge/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-t06-llm-judge/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06-llm-judge/steps/correctness/instruction.md b/evals/team-tasks/vo-t06-llm-judge/steps/correctness/instruction.md index 5e8f3cf..f32810e 100644 --- a/evals/team-tasks/vo-t06-llm-judge/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-t06-llm-judge/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06-llm-judge/steps/language/instruction.md b/evals/team-tasks/vo-t06-llm-judge/steps/language/instruction.md index d306075..b39c273 100644 --- a/evals/team-tasks/vo-t06-llm-judge/steps/language/instruction.md +++ b/evals/team-tasks/vo-t06-llm-judge/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06-llm-judge/steps/security/instruction.md b/evals/team-tasks/vo-t06-llm-judge/steps/security/instruction.md index df805ab..778ed1c 100644 --- a/evals/team-tasks/vo-t06-llm-judge/steps/security/instruction.md +++ b/evals/team-tasks/vo-t06-llm-judge/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06-llm-judge/steps/tests/instruction.md b/evals/team-tasks/vo-t06-llm-judge/steps/tests/instruction.md index 0da921d..e67180c 100644 --- a/evals/team-tasks/vo-t06-llm-judge/steps/tests/instruction.md +++ b/evals/team-tasks/vo-t06-llm-judge/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06-llm-judge/steps/validator/instruction.md b/evals/team-tasks/vo-t06-llm-judge/steps/validator/instruction.md index d101e27..c3c03a0 100644 --- a/evals/team-tasks/vo-t06-llm-judge/steps/validator/instruction.md +++ b/evals/team-tasks/vo-t06-llm-judge/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is 12998e404b8d505056221802db4756d39adbd5a0. +Use the review-validator skill. The output folder is /workspace/review. The base ref is d47ca8270bc57385662f40f0c12f83f92b572c96. diff --git a/evals/team-tasks/vo-t06-llm-judge/steps/validator/tests/workspace/repo.tar.gz b/evals/team-tasks/vo-t06-llm-judge/steps/validator/tests/workspace/repo.tar.gz index 32f5d55..45a686d 100644 Binary files a/evals/team-tasks/vo-t06-llm-judge/steps/validator/tests/workspace/repo.tar.gz and b/evals/team-tasks/vo-t06-llm-judge/steps/validator/tests/workspace/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t06-llm-judge/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-t06-llm-judge/steps/validator/workdir/setup.sh index 60dd6c8..1206573 100755 --- a/evals/team-tasks/vo-t06-llm-judge/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-t06-llm-judge/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 12998e404b8d505056221802db4756d39adbd5a0 --head b27e76080c69c9fe7eb127e1db7f1f9ae692088a +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base d47ca8270bc57385662f40f0c12f83f92b572c96 --head d9677189b6880ce609fd32f9132b8fc965fac999 diff --git a/evals/team-tasks/vo-t06/environment/repo.tar.gz b/evals/team-tasks/vo-t06/environment/repo.tar.gz index 32f5d55..45a686d 100644 Binary files a/evals/team-tasks/vo-t06/environment/repo.tar.gz and b/evals/team-tasks/vo-t06/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-t06/instruction.md b/evals/team-tasks/vo-t06/instruction.md index c468af4..ea5c375 100644 --- a/evals/team-tasks/vo-t06/instruction.md +++ b/evals/team-tasks/vo-t06/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `12998e404b8d505056221802db4756d39adbd5a0` (base) and -`b27e76080c69c9fe7eb127e1db7f1f9ae692088a` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `d47ca8270bc57385662f40f0c12f83f92b572c96` (base) and +`d9677189b6880ce609fd32f9132b8fc965fac999` (head). What the change is meant to do: The change adds region capture to screen recording. After the user picks a screen, a selection overlay shows a captured frame and lets them draw a rectangle, and the recorder crops the stream to that rectangle. It also maps the capture mode to the picker's surface hint, minimizes and restores the app window around the picker, waits for the first video frame before compositing, and sets a dark background color on the window and page. diff --git a/evals/team-tasks/vo-t06/steps/architecture/instruction.md b/evals/team-tasks/vo-t06/steps/architecture/instruction.md index f0a1014..910ad0b 100644 --- a/evals/team-tasks/vo-t06/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-t06/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06/steps/code_quality/instruction.md b/evals/team-tasks/vo-t06/steps/code_quality/instruction.md index 2dfc5aa..efa1969 100644 --- a/evals/team-tasks/vo-t06/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-t06/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06/steps/compliance/instruction.md b/evals/team-tasks/vo-t06/steps/compliance/instruction.md index 3c0c51f..12aab79 100644 --- a/evals/team-tasks/vo-t06/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-t06/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06/steps/correctness/instruction.md b/evals/team-tasks/vo-t06/steps/correctness/instruction.md index 5e8f3cf..f32810e 100644 --- a/evals/team-tasks/vo-t06/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-t06/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06/steps/language/instruction.md b/evals/team-tasks/vo-t06/steps/language/instruction.md index d306075..b39c273 100644 --- a/evals/team-tasks/vo-t06/steps/language/instruction.md +++ b/evals/team-tasks/vo-t06/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06/steps/security/instruction.md b/evals/team-tasks/vo-t06/steps/security/instruction.md index df805ab..778ed1c 100644 --- a/evals/team-tasks/vo-t06/steps/security/instruction.md +++ b/evals/team-tasks/vo-t06/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06/steps/tests/instruction.md b/evals/team-tasks/vo-t06/steps/tests/instruction.md index 0da921d..e67180c 100644 --- a/evals/team-tasks/vo-t06/steps/tests/instruction.md +++ b/evals/team-tasks/vo-t06/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs 12998e404b8d505056221802db4756d39adbd5a0 and b27e76080c69c9fe7eb127e1db7f1f9ae692088a in the repository in /workspace/repo. +Review the change between git refs d47ca8270bc57385662f40f0c12f83f92b572c96 and d9677189b6880ce609fd32f9132b8fc965fac999 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-t06/steps/validator/instruction.md b/evals/team-tasks/vo-t06/steps/validator/instruction.md index d101e27..c3c03a0 100644 --- a/evals/team-tasks/vo-t06/steps/validator/instruction.md +++ b/evals/team-tasks/vo-t06/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is 12998e404b8d505056221802db4756d39adbd5a0. +Use the review-validator skill. The output folder is /workspace/review. The base ref is d47ca8270bc57385662f40f0c12f83f92b572c96. diff --git a/evals/team-tasks/vo-t06/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-t06/steps/validator/workdir/setup.sh index 60dd6c8..1206573 100755 --- a/evals/team-tasks/vo-t06/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-t06/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 12998e404b8d505056221802db4756d39adbd5a0 --head b27e76080c69c9fe7eb127e1db7f1f9ae692088a +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base d47ca8270bc57385662f40f0c12f83f92b572c96 --head d9677189b6880ce609fd32f9132b8fc965fac999 diff --git a/evals/team-tasks/vo-x1/environment/repo.tar.gz b/evals/team-tasks/vo-x1/environment/repo.tar.gz index 4f24454..c10c3ac 100644 Binary files a/evals/team-tasks/vo-x1/environment/repo.tar.gz and b/evals/team-tasks/vo-x1/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-x1/instruction.md b/evals/team-tasks/vo-x1/instruction.md index ceb7e8a..85f7c9b 100644 --- a/evals/team-tasks/vo-x1/instruction.md +++ b/evals/team-tasks/vo-x1/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `12cf2440862b1d9b3d5221eedfe30161279aba17` (base) and -`6d2d586d7ab769f5c27e8b48fa817687b265ac28` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `02ac4d2949831e10616c378437beaf866f626f26` (base) and +`1300764111acc28a5c4b093d6577116392173f88` (head). What the change is meant to do: The change adds a session-id check to the recording commands, an expiry timestamp for Drive tokens, a timeout on the Drive sign-in callback, async video-tool calls and a content security policy in the window config. The frontend shares one environment-detection helper instead of local copies, sends recording chunks as byte arrays, listens for widget stop and cancel events on the home page and refreshes expired Drive tokens before upload. It removes two unused Rust dependencies and blanks credentials in the config file synced to the static folder. diff --git a/evals/team-tasks/vo-x1/steps/architecture/instruction.md b/evals/team-tasks/vo-x1/steps/architecture/instruction.md index c014439..184a40c 100644 --- a/evals/team-tasks/vo-x1/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-x1/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs 12cf2440862b1d9b3d5221eedfe30161279aba17 and 6d2d586d7ab769f5c27e8b48fa817687b265ac28 in the repository in /workspace/repo. +Review the change between git refs 02ac4d2949831e10616c378437beaf866f626f26 and 1300764111acc28a5c4b093d6577116392173f88 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x1/steps/code_quality/instruction.md b/evals/team-tasks/vo-x1/steps/code_quality/instruction.md index b5ef78a..cd00584 100644 --- a/evals/team-tasks/vo-x1/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-x1/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs 12cf2440862b1d9b3d5221eedfe30161279aba17 and 6d2d586d7ab769f5c27e8b48fa817687b265ac28 in the repository in /workspace/repo. +Review the change between git refs 02ac4d2949831e10616c378437beaf866f626f26 and 1300764111acc28a5c4b093d6577116392173f88 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x1/steps/compliance/instruction.md b/evals/team-tasks/vo-x1/steps/compliance/instruction.md index f677667..bf851f6 100644 --- a/evals/team-tasks/vo-x1/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-x1/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs 12cf2440862b1d9b3d5221eedfe30161279aba17 and 6d2d586d7ab769f5c27e8b48fa817687b265ac28 in the repository in /workspace/repo. +Review the change between git refs 02ac4d2949831e10616c378437beaf866f626f26 and 1300764111acc28a5c4b093d6577116392173f88 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x1/steps/correctness/instruction.md b/evals/team-tasks/vo-x1/steps/correctness/instruction.md index 6ccc94d..aa34a40 100644 --- a/evals/team-tasks/vo-x1/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-x1/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs 12cf2440862b1d9b3d5221eedfe30161279aba17 and 6d2d586d7ab769f5c27e8b48fa817687b265ac28 in the repository in /workspace/repo. +Review the change between git refs 02ac4d2949831e10616c378437beaf866f626f26 and 1300764111acc28a5c4b093d6577116392173f88 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x1/steps/language/instruction.md b/evals/team-tasks/vo-x1/steps/language/instruction.md index 6b5fcbd..7b9f8f9 100644 --- a/evals/team-tasks/vo-x1/steps/language/instruction.md +++ b/evals/team-tasks/vo-x1/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs 12cf2440862b1d9b3d5221eedfe30161279aba17 and 6d2d586d7ab769f5c27e8b48fa817687b265ac28 in the repository in /workspace/repo. +Review the change between git refs 02ac4d2949831e10616c378437beaf866f626f26 and 1300764111acc28a5c4b093d6577116392173f88 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x1/steps/security/instruction.md b/evals/team-tasks/vo-x1/steps/security/instruction.md index 9908da4..0e6dfdd 100644 --- a/evals/team-tasks/vo-x1/steps/security/instruction.md +++ b/evals/team-tasks/vo-x1/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs 12cf2440862b1d9b3d5221eedfe30161279aba17 and 6d2d586d7ab769f5c27e8b48fa817687b265ac28 in the repository in /workspace/repo. +Review the change between git refs 02ac4d2949831e10616c378437beaf866f626f26 and 1300764111acc28a5c4b093d6577116392173f88 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x1/steps/tests/instruction.md b/evals/team-tasks/vo-x1/steps/tests/instruction.md index 3a5a625..26475f8 100644 --- a/evals/team-tasks/vo-x1/steps/tests/instruction.md +++ b/evals/team-tasks/vo-x1/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs 12cf2440862b1d9b3d5221eedfe30161279aba17 and 6d2d586d7ab769f5c27e8b48fa817687b265ac28 in the repository in /workspace/repo. +Review the change between git refs 02ac4d2949831e10616c378437beaf866f626f26 and 1300764111acc28a5c4b093d6577116392173f88 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x1/steps/validator/instruction.md b/evals/team-tasks/vo-x1/steps/validator/instruction.md index e1f93df..e5be68a 100644 --- a/evals/team-tasks/vo-x1/steps/validator/instruction.md +++ b/evals/team-tasks/vo-x1/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is 12cf2440862b1d9b3d5221eedfe30161279aba17. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 02ac4d2949831e10616c378437beaf866f626f26. diff --git a/evals/team-tasks/vo-x1/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-x1/steps/validator/workdir/setup.sh index 18c51e5..8726e06 100755 --- a/evals/team-tasks/vo-x1/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-x1/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 12cf2440862b1d9b3d5221eedfe30161279aba17 --head 6d2d586d7ab769f5c27e8b48fa817687b265ac28 +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 02ac4d2949831e10616c378437beaf866f626f26 --head 1300764111acc28a5c4b093d6577116392173f88 diff --git a/evals/team-tasks/vo-x17/environment/repo.tar.gz b/evals/team-tasks/vo-x17/environment/repo.tar.gz index cbdc463..94839e9 100644 Binary files a/evals/team-tasks/vo-x17/environment/repo.tar.gz and b/evals/team-tasks/vo-x17/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-x17/instruction.md b/evals/team-tasks/vo-x17/instruction.md index 7520b02..a7614a1 100644 --- a/evals/team-tasks/vo-x17/instruction.md +++ b/evals/team-tasks/vo-x17/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `a6a844f9815f8c2ce9d6dff6ac76e596ba2e0a11` (base) and -`b8cbb3891ad04bb2f740a419ed002ee97ef3ea08` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `21d799db2caf42be15514065780e0faadaae83d9` (base) and +`bb090a4a9bf3951afa70d85e4bb727948eeba754` (head). What the change is meant to do: The change reorganises sidecar HTTP and generation code and routes generated audio through the desktop backend. The Rust side registers one shared HTTP client with per-request timeouts, adds a command that returns generated audio bytes, validates generation ids and waits for sidecar startup by notification instead of polling. The Python server moves generation and voice conversion into separate functions run by the queue, and common page styles move into the global stylesheet. The content security policy no longer allows media from localhost. diff --git a/evals/team-tasks/vo-x17/steps/architecture/instruction.md b/evals/team-tasks/vo-x17/steps/architecture/instruction.md index a6769be..2a80f58 100644 --- a/evals/team-tasks/vo-x17/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-x17/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs a6a844f9815f8c2ce9d6dff6ac76e596ba2e0a11 and b8cbb3891ad04bb2f740a419ed002ee97ef3ea08 in the repository in /workspace/repo. +Review the change between git refs 21d799db2caf42be15514065780e0faadaae83d9 and bb090a4a9bf3951afa70d85e4bb727948eeba754 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x17/steps/code_quality/instruction.md b/evals/team-tasks/vo-x17/steps/code_quality/instruction.md index e82e7f0..bdc49e0 100644 --- a/evals/team-tasks/vo-x17/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-x17/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs a6a844f9815f8c2ce9d6dff6ac76e596ba2e0a11 and b8cbb3891ad04bb2f740a419ed002ee97ef3ea08 in the repository in /workspace/repo. +Review the change between git refs 21d799db2caf42be15514065780e0faadaae83d9 and bb090a4a9bf3951afa70d85e4bb727948eeba754 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x17/steps/compliance/instruction.md b/evals/team-tasks/vo-x17/steps/compliance/instruction.md index 561639e..09b1366 100644 --- a/evals/team-tasks/vo-x17/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-x17/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs a6a844f9815f8c2ce9d6dff6ac76e596ba2e0a11 and b8cbb3891ad04bb2f740a419ed002ee97ef3ea08 in the repository in /workspace/repo. +Review the change between git refs 21d799db2caf42be15514065780e0faadaae83d9 and bb090a4a9bf3951afa70d85e4bb727948eeba754 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x17/steps/correctness/instruction.md b/evals/team-tasks/vo-x17/steps/correctness/instruction.md index 1334d27..efef29e 100644 --- a/evals/team-tasks/vo-x17/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-x17/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs a6a844f9815f8c2ce9d6dff6ac76e596ba2e0a11 and b8cbb3891ad04bb2f740a419ed002ee97ef3ea08 in the repository in /workspace/repo. +Review the change between git refs 21d799db2caf42be15514065780e0faadaae83d9 and bb090a4a9bf3951afa70d85e4bb727948eeba754 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x17/steps/language/instruction.md b/evals/team-tasks/vo-x17/steps/language/instruction.md index a4c71b8..daec874 100644 --- a/evals/team-tasks/vo-x17/steps/language/instruction.md +++ b/evals/team-tasks/vo-x17/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs a6a844f9815f8c2ce9d6dff6ac76e596ba2e0a11 and b8cbb3891ad04bb2f740a419ed002ee97ef3ea08 in the repository in /workspace/repo. +Review the change between git refs 21d799db2caf42be15514065780e0faadaae83d9 and bb090a4a9bf3951afa70d85e4bb727948eeba754 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x17/steps/security/instruction.md b/evals/team-tasks/vo-x17/steps/security/instruction.md index e90208f..56cdf91 100644 --- a/evals/team-tasks/vo-x17/steps/security/instruction.md +++ b/evals/team-tasks/vo-x17/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs a6a844f9815f8c2ce9d6dff6ac76e596ba2e0a11 and b8cbb3891ad04bb2f740a419ed002ee97ef3ea08 in the repository in /workspace/repo. +Review the change between git refs 21d799db2caf42be15514065780e0faadaae83d9 and bb090a4a9bf3951afa70d85e4bb727948eeba754 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x17/steps/tests/instruction.md b/evals/team-tasks/vo-x17/steps/tests/instruction.md index dccf123..c39760e 100644 --- a/evals/team-tasks/vo-x17/steps/tests/instruction.md +++ b/evals/team-tasks/vo-x17/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs a6a844f9815f8c2ce9d6dff6ac76e596ba2e0a11 and b8cbb3891ad04bb2f740a419ed002ee97ef3ea08 in the repository in /workspace/repo. +Review the change between git refs 21d799db2caf42be15514065780e0faadaae83d9 and bb090a4a9bf3951afa70d85e4bb727948eeba754 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x17/steps/validator/instruction.md b/evals/team-tasks/vo-x17/steps/validator/instruction.md index be1196e..07c3428 100644 --- a/evals/team-tasks/vo-x17/steps/validator/instruction.md +++ b/evals/team-tasks/vo-x17/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is a6a844f9815f8c2ce9d6dff6ac76e596ba2e0a11. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 21d799db2caf42be15514065780e0faadaae83d9. diff --git a/evals/team-tasks/vo-x17/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-x17/steps/validator/workdir/setup.sh index 5a28263..06b765b 100755 --- a/evals/team-tasks/vo-x17/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-x17/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base a6a844f9815f8c2ce9d6dff6ac76e596ba2e0a11 --head b8cbb3891ad04bb2f740a419ed002ee97ef3ea08 +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 21d799db2caf42be15514065780e0faadaae83d9 --head bb090a4a9bf3951afa70d85e4bb727948eeba754 diff --git a/evals/team-tasks/vo-x20/environment/repo.tar.gz b/evals/team-tasks/vo-x20/environment/repo.tar.gz index 56581c2..397b0ae 100644 Binary files a/evals/team-tasks/vo-x20/environment/repo.tar.gz and b/evals/team-tasks/vo-x20/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-x20/instruction.md b/evals/team-tasks/vo-x20/instruction.md index a58feec..9b90917 100644 --- a/evals/team-tasks/vo-x20/instruction.md +++ b/evals/team-tasks/vo-x20/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `f3d8d6db39233016be1ef4f0c08b9993bea4f28d` (base) and -`47f76c29d4bbf66d2df8ca04929894bfc698882d` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `1bba865ce5572776ba281dcefae126d4b74ed0fd` (base) and +`b3b3207e612681e8a48c0ba284ac2a048fc56572` (head). What the change is meant to do: The change lets recordings be renamed from the video library, renaming every companion file, and reworks text-to-speech timing: audio duration probing, transcription accuracy and the repair of word timestamps. It prevents audio clipping in both text-to-speech and voice conversion modes, adds the CosyVoice3 model to the prerequisites and model downloads, signs the sidecar build, reads the ElevenLabs API key from the environment when it is set, and updates the README. diff --git a/evals/team-tasks/vo-x20/steps/architecture/instruction.md b/evals/team-tasks/vo-x20/steps/architecture/instruction.md index caff4fe..8151d1a 100644 --- a/evals/team-tasks/vo-x20/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-x20/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs f3d8d6db39233016be1ef4f0c08b9993bea4f28d and 47f76c29d4bbf66d2df8ca04929894bfc698882d in the repository in /workspace/repo. +Review the change between git refs 1bba865ce5572776ba281dcefae126d4b74ed0fd and b3b3207e612681e8a48c0ba284ac2a048fc56572 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x20/steps/code_quality/instruction.md b/evals/team-tasks/vo-x20/steps/code_quality/instruction.md index d5297cd..fd8b958 100644 --- a/evals/team-tasks/vo-x20/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-x20/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs f3d8d6db39233016be1ef4f0c08b9993bea4f28d and 47f76c29d4bbf66d2df8ca04929894bfc698882d in the repository in /workspace/repo. +Review the change between git refs 1bba865ce5572776ba281dcefae126d4b74ed0fd and b3b3207e612681e8a48c0ba284ac2a048fc56572 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x20/steps/compliance/instruction.md b/evals/team-tasks/vo-x20/steps/compliance/instruction.md index aca5d39..f772600 100644 --- a/evals/team-tasks/vo-x20/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-x20/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs f3d8d6db39233016be1ef4f0c08b9993bea4f28d and 47f76c29d4bbf66d2df8ca04929894bfc698882d in the repository in /workspace/repo. +Review the change between git refs 1bba865ce5572776ba281dcefae126d4b74ed0fd and b3b3207e612681e8a48c0ba284ac2a048fc56572 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x20/steps/correctness/instruction.md b/evals/team-tasks/vo-x20/steps/correctness/instruction.md index db1606e..5ee833c 100644 --- a/evals/team-tasks/vo-x20/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-x20/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs f3d8d6db39233016be1ef4f0c08b9993bea4f28d and 47f76c29d4bbf66d2df8ca04929894bfc698882d in the repository in /workspace/repo. +Review the change between git refs 1bba865ce5572776ba281dcefae126d4b74ed0fd and b3b3207e612681e8a48c0ba284ac2a048fc56572 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x20/steps/language/instruction.md b/evals/team-tasks/vo-x20/steps/language/instruction.md index d7ff7a6..22a87fc 100644 --- a/evals/team-tasks/vo-x20/steps/language/instruction.md +++ b/evals/team-tasks/vo-x20/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs f3d8d6db39233016be1ef4f0c08b9993bea4f28d and 47f76c29d4bbf66d2df8ca04929894bfc698882d in the repository in /workspace/repo. +Review the change between git refs 1bba865ce5572776ba281dcefae126d4b74ed0fd and b3b3207e612681e8a48c0ba284ac2a048fc56572 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x20/steps/security/instruction.md b/evals/team-tasks/vo-x20/steps/security/instruction.md index 6225bc6..a59e70a 100644 --- a/evals/team-tasks/vo-x20/steps/security/instruction.md +++ b/evals/team-tasks/vo-x20/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs f3d8d6db39233016be1ef4f0c08b9993bea4f28d and 47f76c29d4bbf66d2df8ca04929894bfc698882d in the repository in /workspace/repo. +Review the change between git refs 1bba865ce5572776ba281dcefae126d4b74ed0fd and b3b3207e612681e8a48c0ba284ac2a048fc56572 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x20/steps/tests/instruction.md b/evals/team-tasks/vo-x20/steps/tests/instruction.md index 6399597..e8d5e83 100644 --- a/evals/team-tasks/vo-x20/steps/tests/instruction.md +++ b/evals/team-tasks/vo-x20/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs f3d8d6db39233016be1ef4f0c08b9993bea4f28d and 47f76c29d4bbf66d2df8ca04929894bfc698882d in the repository in /workspace/repo. +Review the change between git refs 1bba865ce5572776ba281dcefae126d4b74ed0fd and b3b3207e612681e8a48c0ba284ac2a048fc56572 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x20/steps/validator/instruction.md b/evals/team-tasks/vo-x20/steps/validator/instruction.md index b3c0f69..8cd301e 100644 --- a/evals/team-tasks/vo-x20/steps/validator/instruction.md +++ b/evals/team-tasks/vo-x20/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is f3d8d6db39233016be1ef4f0c08b9993bea4f28d. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 1bba865ce5572776ba281dcefae126d4b74ed0fd. diff --git a/evals/team-tasks/vo-x20/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-x20/steps/validator/workdir/setup.sh index 38761c4..e5a11b6 100755 --- a/evals/team-tasks/vo-x20/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-x20/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base f3d8d6db39233016be1ef4f0c08b9993bea4f28d --head 47f76c29d4bbf66d2df8ca04929894bfc698882d +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 1bba865ce5572776ba281dcefae126d4b74ed0fd --head b3b3207e612681e8a48c0ba284ac2a048fc56572 diff --git a/evals/team-tasks/vo-x4-llm-judge/environment/repo.tar.gz b/evals/team-tasks/vo-x4-llm-judge/environment/repo.tar.gz index dbb1d4f..da5559b 100644 Binary files a/evals/team-tasks/vo-x4-llm-judge/environment/repo.tar.gz and b/evals/team-tasks/vo-x4-llm-judge/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-x4-llm-judge/instruction.md b/evals/team-tasks/vo-x4-llm-judge/instruction.md index a46ce0a..36024e2 100644 --- a/evals/team-tasks/vo-x4-llm-judge/instruction.md +++ b/evals/team-tasks/vo-x4-llm-judge/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `b957febf3680dbc91ad08b67333b1ff649f997d5` (base) and -`3a0594941a1ae4200d6d532be5ae1a99b58f0201` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `966a2d8778fe9a9cadf11431478349771365820f` (base) and +`02b02c9de77c22a11c5e804064acffaf3c326fed` (head). What the change is meant to do: The change lets the app seed its configuration from a project-level static config file when no saved config exists. The Rust config loader reads that file from the project root if present and writes it to app data, and the sync to the static folder is skipped when no API key is set. On the frontend, the loader accepts a config only if it includes an API key, falls back to the static file otherwise, and saves that config back to app data in the desktop app. diff --git a/evals/team-tasks/vo-x4-llm-judge/steps/architecture/instruction.md b/evals/team-tasks/vo-x4-llm-judge/steps/architecture/instruction.md index ebadd98..6decaf9 100644 --- a/evals/team-tasks/vo-x4-llm-judge/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-x4-llm-judge/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4-llm-judge/steps/code_quality/instruction.md b/evals/team-tasks/vo-x4-llm-judge/steps/code_quality/instruction.md index 6411cba..1ee8461 100644 --- a/evals/team-tasks/vo-x4-llm-judge/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-x4-llm-judge/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4-llm-judge/steps/compliance/instruction.md b/evals/team-tasks/vo-x4-llm-judge/steps/compliance/instruction.md index ca88dbe..d2ef86d 100644 --- a/evals/team-tasks/vo-x4-llm-judge/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-x4-llm-judge/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4-llm-judge/steps/correctness/instruction.md b/evals/team-tasks/vo-x4-llm-judge/steps/correctness/instruction.md index 453c82c..b52cbdf 100644 --- a/evals/team-tasks/vo-x4-llm-judge/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-x4-llm-judge/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4-llm-judge/steps/language/instruction.md b/evals/team-tasks/vo-x4-llm-judge/steps/language/instruction.md index faeb9f0..6373543 100644 --- a/evals/team-tasks/vo-x4-llm-judge/steps/language/instruction.md +++ b/evals/team-tasks/vo-x4-llm-judge/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4-llm-judge/steps/security/instruction.md b/evals/team-tasks/vo-x4-llm-judge/steps/security/instruction.md index cbe9bab..89f8458 100644 --- a/evals/team-tasks/vo-x4-llm-judge/steps/security/instruction.md +++ b/evals/team-tasks/vo-x4-llm-judge/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4-llm-judge/steps/tests/instruction.md b/evals/team-tasks/vo-x4-llm-judge/steps/tests/instruction.md index da9f70e..a726fb5 100644 --- a/evals/team-tasks/vo-x4-llm-judge/steps/tests/instruction.md +++ b/evals/team-tasks/vo-x4-llm-judge/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4-llm-judge/steps/validator/instruction.md b/evals/team-tasks/vo-x4-llm-judge/steps/validator/instruction.md index ea6bbf1..e060eca 100644 --- a/evals/team-tasks/vo-x4-llm-judge/steps/validator/instruction.md +++ b/evals/team-tasks/vo-x4-llm-judge/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is b957febf3680dbc91ad08b67333b1ff649f997d5. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 966a2d8778fe9a9cadf11431478349771365820f. diff --git a/evals/team-tasks/vo-x4-llm-judge/steps/validator/tests/workspace/repo.tar.gz b/evals/team-tasks/vo-x4-llm-judge/steps/validator/tests/workspace/repo.tar.gz index dbb1d4f..da5559b 100644 Binary files a/evals/team-tasks/vo-x4-llm-judge/steps/validator/tests/workspace/repo.tar.gz and b/evals/team-tasks/vo-x4-llm-judge/steps/validator/tests/workspace/repo.tar.gz differ diff --git a/evals/team-tasks/vo-x4-llm-judge/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-x4-llm-judge/steps/validator/workdir/setup.sh index 5d95828..395b507 100755 --- a/evals/team-tasks/vo-x4-llm-judge/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-x4-llm-judge/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base b957febf3680dbc91ad08b67333b1ff649f997d5 --head 3a0594941a1ae4200d6d532be5ae1a99b58f0201 +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 966a2d8778fe9a9cadf11431478349771365820f --head 02b02c9de77c22a11c5e804064acffaf3c326fed diff --git a/evals/team-tasks/vo-x4/environment/repo.tar.gz b/evals/team-tasks/vo-x4/environment/repo.tar.gz index dbb1d4f..da5559b 100644 Binary files a/evals/team-tasks/vo-x4/environment/repo.tar.gz and b/evals/team-tasks/vo-x4/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-x4/instruction.md b/evals/team-tasks/vo-x4/instruction.md index a46ce0a..36024e2 100644 --- a/evals/team-tasks/vo-x4/instruction.md +++ b/evals/team-tasks/vo-x4/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `b957febf3680dbc91ad08b67333b1ff649f997d5` (base) and -`3a0594941a1ae4200d6d532be5ae1a99b58f0201` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `966a2d8778fe9a9cadf11431478349771365820f` (base) and +`02b02c9de77c22a11c5e804064acffaf3c326fed` (head). What the change is meant to do: The change lets the app seed its configuration from a project-level static config file when no saved config exists. The Rust config loader reads that file from the project root if present and writes it to app data, and the sync to the static folder is skipped when no API key is set. On the frontend, the loader accepts a config only if it includes an API key, falls back to the static file otherwise, and saves that config back to app data in the desktop app. diff --git a/evals/team-tasks/vo-x4/steps/architecture/instruction.md b/evals/team-tasks/vo-x4/steps/architecture/instruction.md index ebadd98..6decaf9 100644 --- a/evals/team-tasks/vo-x4/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-x4/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4/steps/code_quality/instruction.md b/evals/team-tasks/vo-x4/steps/code_quality/instruction.md index 6411cba..1ee8461 100644 --- a/evals/team-tasks/vo-x4/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-x4/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4/steps/compliance/instruction.md b/evals/team-tasks/vo-x4/steps/compliance/instruction.md index ca88dbe..d2ef86d 100644 --- a/evals/team-tasks/vo-x4/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-x4/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4/steps/correctness/instruction.md b/evals/team-tasks/vo-x4/steps/correctness/instruction.md index 453c82c..b52cbdf 100644 --- a/evals/team-tasks/vo-x4/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-x4/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4/steps/language/instruction.md b/evals/team-tasks/vo-x4/steps/language/instruction.md index faeb9f0..6373543 100644 --- a/evals/team-tasks/vo-x4/steps/language/instruction.md +++ b/evals/team-tasks/vo-x4/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4/steps/security/instruction.md b/evals/team-tasks/vo-x4/steps/security/instruction.md index cbe9bab..89f8458 100644 --- a/evals/team-tasks/vo-x4/steps/security/instruction.md +++ b/evals/team-tasks/vo-x4/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4/steps/tests/instruction.md b/evals/team-tasks/vo-x4/steps/tests/instruction.md index da9f70e..a726fb5 100644 --- a/evals/team-tasks/vo-x4/steps/tests/instruction.md +++ b/evals/team-tasks/vo-x4/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs b957febf3680dbc91ad08b67333b1ff649f997d5 and 3a0594941a1ae4200d6d532be5ae1a99b58f0201 in the repository in /workspace/repo. +Review the change between git refs 966a2d8778fe9a9cadf11431478349771365820f and 02b02c9de77c22a11c5e804064acffaf3c326fed in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x4/steps/validator/instruction.md b/evals/team-tasks/vo-x4/steps/validator/instruction.md index ea6bbf1..e060eca 100644 --- a/evals/team-tasks/vo-x4/steps/validator/instruction.md +++ b/evals/team-tasks/vo-x4/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is b957febf3680dbc91ad08b67333b1ff649f997d5. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 966a2d8778fe9a9cadf11431478349771365820f. diff --git a/evals/team-tasks/vo-x4/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-x4/steps/validator/workdir/setup.sh index 5d95828..395b507 100755 --- a/evals/team-tasks/vo-x4/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-x4/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base b957febf3680dbc91ad08b67333b1ff649f997d5 --head 3a0594941a1ae4200d6d532be5ae1a99b58f0201 +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 966a2d8778fe9a9cadf11431478349771365820f --head 02b02c9de77c22a11c5e804064acffaf3c326fed diff --git a/evals/team-tasks/vo-x7-llm-judge/environment/repo.tar.gz b/evals/team-tasks/vo-x7-llm-judge/environment/repo.tar.gz index 3dccca2..31d2c69 100644 Binary files a/evals/team-tasks/vo-x7-llm-judge/environment/repo.tar.gz and b/evals/team-tasks/vo-x7-llm-judge/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-x7-llm-judge/instruction.md b/evals/team-tasks/vo-x7-llm-judge/instruction.md index 140bb66..f116885 100644 --- a/evals/team-tasks/vo-x7-llm-judge/instruction.md +++ b/evals/team-tasks/vo-x7-llm-judge/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `748d85a9a8cd04b0d38039ddd2489363d3971528` (base) and -`a66bd3b1e89d8e4d07d8991caa9f41f00a952f28` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `1c236b38a1908ac0957466bbbb94dcedfed3ce20` (base) and +`4e456c8999479b805aaa7bedec6a0a25327bb481` (head). What the change is meant to do: The change simplifies how configuration is synced to the static folder and how Drive tokens are refreshed before upload. The Rust sync writes the full configuration instead of a reduced copy with credentials blanked, and its doc comment notes the file is gitignored. The preview page checks token expiry once before upload, for both desktop and browser modes, and refreshes the token when the expiry is missing or has passed. diff --git a/evals/team-tasks/vo-x7-llm-judge/steps/architecture/instruction.md b/evals/team-tasks/vo-x7-llm-judge/steps/architecture/instruction.md index fdb50c5..c1f3413 100644 --- a/evals/team-tasks/vo-x7-llm-judge/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-x7-llm-judge/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7-llm-judge/steps/code_quality/instruction.md b/evals/team-tasks/vo-x7-llm-judge/steps/code_quality/instruction.md index d00ab02..b6adcf9 100644 --- a/evals/team-tasks/vo-x7-llm-judge/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-x7-llm-judge/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7-llm-judge/steps/compliance/instruction.md b/evals/team-tasks/vo-x7-llm-judge/steps/compliance/instruction.md index 3dd9673..d920942 100644 --- a/evals/team-tasks/vo-x7-llm-judge/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-x7-llm-judge/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7-llm-judge/steps/correctness/instruction.md b/evals/team-tasks/vo-x7-llm-judge/steps/correctness/instruction.md index 8356b72..9b8e350 100644 --- a/evals/team-tasks/vo-x7-llm-judge/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-x7-llm-judge/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7-llm-judge/steps/language/instruction.md b/evals/team-tasks/vo-x7-llm-judge/steps/language/instruction.md index 4011256..36802f3 100644 --- a/evals/team-tasks/vo-x7-llm-judge/steps/language/instruction.md +++ b/evals/team-tasks/vo-x7-llm-judge/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7-llm-judge/steps/security/instruction.md b/evals/team-tasks/vo-x7-llm-judge/steps/security/instruction.md index 5c21dd9..5fb17ba 100644 --- a/evals/team-tasks/vo-x7-llm-judge/steps/security/instruction.md +++ b/evals/team-tasks/vo-x7-llm-judge/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7-llm-judge/steps/tests/instruction.md b/evals/team-tasks/vo-x7-llm-judge/steps/tests/instruction.md index dd11396..f654d65 100644 --- a/evals/team-tasks/vo-x7-llm-judge/steps/tests/instruction.md +++ b/evals/team-tasks/vo-x7-llm-judge/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7-llm-judge/steps/validator/instruction.md b/evals/team-tasks/vo-x7-llm-judge/steps/validator/instruction.md index b00e722..d605cd6 100644 --- a/evals/team-tasks/vo-x7-llm-judge/steps/validator/instruction.md +++ b/evals/team-tasks/vo-x7-llm-judge/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is 748d85a9a8cd04b0d38039ddd2489363d3971528. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 1c236b38a1908ac0957466bbbb94dcedfed3ce20. diff --git a/evals/team-tasks/vo-x7-llm-judge/steps/validator/tests/workspace/repo.tar.gz b/evals/team-tasks/vo-x7-llm-judge/steps/validator/tests/workspace/repo.tar.gz index 3dccca2..31d2c69 100644 Binary files a/evals/team-tasks/vo-x7-llm-judge/steps/validator/tests/workspace/repo.tar.gz and b/evals/team-tasks/vo-x7-llm-judge/steps/validator/tests/workspace/repo.tar.gz differ diff --git a/evals/team-tasks/vo-x7-llm-judge/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-x7-llm-judge/steps/validator/workdir/setup.sh index 76153e4..12f617e 100755 --- a/evals/team-tasks/vo-x7-llm-judge/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-x7-llm-judge/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 748d85a9a8cd04b0d38039ddd2489363d3971528 --head a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 1c236b38a1908ac0957466bbbb94dcedfed3ce20 --head 4e456c8999479b805aaa7bedec6a0a25327bb481 diff --git a/evals/team-tasks/vo-x7/environment/repo.tar.gz b/evals/team-tasks/vo-x7/environment/repo.tar.gz index 3dccca2..31d2c69 100644 Binary files a/evals/team-tasks/vo-x7/environment/repo.tar.gz and b/evals/team-tasks/vo-x7/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-x7/instruction.md b/evals/team-tasks/vo-x7/instruction.md index 140bb66..f116885 100644 --- a/evals/team-tasks/vo-x7/instruction.md +++ b/evals/team-tasks/vo-x7/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `748d85a9a8cd04b0d38039ddd2489363d3971528` (base) and -`a66bd3b1e89d8e4d07d8991caa9f41f00a952f28` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `1c236b38a1908ac0957466bbbb94dcedfed3ce20` (base) and +`4e456c8999479b805aaa7bedec6a0a25327bb481` (head). What the change is meant to do: The change simplifies how configuration is synced to the static folder and how Drive tokens are refreshed before upload. The Rust sync writes the full configuration instead of a reduced copy with credentials blanked, and its doc comment notes the file is gitignored. The preview page checks token expiry once before upload, for both desktop and browser modes, and refreshes the token when the expiry is missing or has passed. diff --git a/evals/team-tasks/vo-x7/steps/architecture/instruction.md b/evals/team-tasks/vo-x7/steps/architecture/instruction.md index fdb50c5..c1f3413 100644 --- a/evals/team-tasks/vo-x7/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-x7/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7/steps/code_quality/instruction.md b/evals/team-tasks/vo-x7/steps/code_quality/instruction.md index d00ab02..b6adcf9 100644 --- a/evals/team-tasks/vo-x7/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-x7/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7/steps/compliance/instruction.md b/evals/team-tasks/vo-x7/steps/compliance/instruction.md index 3dd9673..d920942 100644 --- a/evals/team-tasks/vo-x7/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-x7/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7/steps/correctness/instruction.md b/evals/team-tasks/vo-x7/steps/correctness/instruction.md index 8356b72..9b8e350 100644 --- a/evals/team-tasks/vo-x7/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-x7/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7/steps/language/instruction.md b/evals/team-tasks/vo-x7/steps/language/instruction.md index 4011256..36802f3 100644 --- a/evals/team-tasks/vo-x7/steps/language/instruction.md +++ b/evals/team-tasks/vo-x7/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7/steps/security/instruction.md b/evals/team-tasks/vo-x7/steps/security/instruction.md index 5c21dd9..5fb17ba 100644 --- a/evals/team-tasks/vo-x7/steps/security/instruction.md +++ b/evals/team-tasks/vo-x7/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7/steps/tests/instruction.md b/evals/team-tasks/vo-x7/steps/tests/instruction.md index dd11396..f654d65 100644 --- a/evals/team-tasks/vo-x7/steps/tests/instruction.md +++ b/evals/team-tasks/vo-x7/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs 748d85a9a8cd04b0d38039ddd2489363d3971528 and a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 in the repository in /workspace/repo. +Review the change between git refs 1c236b38a1908ac0957466bbbb94dcedfed3ce20 and 4e456c8999479b805aaa7bedec6a0a25327bb481 in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x7/steps/validator/instruction.md b/evals/team-tasks/vo-x7/steps/validator/instruction.md index b00e722..d605cd6 100644 --- a/evals/team-tasks/vo-x7/steps/validator/instruction.md +++ b/evals/team-tasks/vo-x7/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is 748d85a9a8cd04b0d38039ddd2489363d3971528. +Use the review-validator skill. The output folder is /workspace/review. The base ref is 1c236b38a1908ac0957466bbbb94dcedfed3ce20. diff --git a/evals/team-tasks/vo-x7/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-x7/steps/validator/workdir/setup.sh index 76153e4..12f617e 100755 --- a/evals/team-tasks/vo-x7/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-x7/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 748d85a9a8cd04b0d38039ddd2489363d3971528 --head a66bd3b1e89d8e4d07d8991caa9f41f00a952f28 +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base 1c236b38a1908ac0957466bbbb94dcedfed3ce20 --head 4e456c8999479b805aaa7bedec6a0a25327bb481 diff --git a/evals/team-tasks/vo-x8/environment/repo.tar.gz b/evals/team-tasks/vo-x8/environment/repo.tar.gz index ba4db29..e1b20e7 100644 Binary files a/evals/team-tasks/vo-x8/environment/repo.tar.gz and b/evals/team-tasks/vo-x8/environment/repo.tar.gz differ diff --git a/evals/team-tasks/vo-x8/instruction.md b/evals/team-tasks/vo-x8/instruction.md index 9d24998..41136a5 100644 --- a/evals/team-tasks/vo-x8/instruction.md +++ b/evals/team-tasks/vo-x8/instruction.md @@ -2,8 +2,8 @@ This is a multi-step task. Each agent reads only its own step's `instruction.md`. -The git repository at `/workspace/repo` holds the change: `d5d5f00f761c3ba95294d72aae04380d7b91ba80` (base) and -`cdac7cc610b5f82199f8d6773921842f92c674b6` (head). What the change is meant to do: +The git repository at `/workspace/repo` holds the change: `ea652265cd8db98f0a52d601325af6de837357b9` (base) and +`5794f727be0b1a2474ac9659058a5efae840a94c` (head). What the change is meant to do: The change makes the Drive link on the preview page open through a button and adjusts video encoding for odd dimensions. The link opens through the shell plugin in the desktop app and in a new tab in the browser, with matching styles. The two commands that re-encode video now pad width and height up to even numbers, and the notes folder is added to gitignore. diff --git a/evals/team-tasks/vo-x8/steps/architecture/instruction.md b/evals/team-tasks/vo-x8/steps/architecture/instruction.md index 65db686..e450626 100644 --- a/evals/team-tasks/vo-x8/steps/architecture/instruction.md +++ b/evals/team-tasks/vo-x8/steps/architecture/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-architecture skill. You are the `architecture` reviewer. -Review the change between git refs d5d5f00f761c3ba95294d72aae04380d7b91ba80 and cdac7cc610b5f82199f8d6773921842f92c674b6 in the repository in /workspace/repo. +Review the change between git refs ea652265cd8db98f0a52d601325af6de837357b9 and 5794f727be0b1a2474ac9659058a5efae840a94c in the repository in /workspace/repo. Write your findings to /workspace/review/findings.architecture.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x8/steps/code_quality/instruction.md b/evals/team-tasks/vo-x8/steps/code_quality/instruction.md index 64e4ff8..c8fb00e 100644 --- a/evals/team-tasks/vo-x8/steps/code_quality/instruction.md +++ b/evals/team-tasks/vo-x8/steps/code_quality/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-code-quality skill. You are the `code_quality` reviewer. -Review the change between git refs d5d5f00f761c3ba95294d72aae04380d7b91ba80 and cdac7cc610b5f82199f8d6773921842f92c674b6 in the repository in /workspace/repo. +Review the change between git refs ea652265cd8db98f0a52d601325af6de837357b9 and 5794f727be0b1a2474ac9659058a5efae840a94c in the repository in /workspace/repo. Write your findings to /workspace/review/findings.code_quality.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x8/steps/compliance/instruction.md b/evals/team-tasks/vo-x8/steps/compliance/instruction.md index ead9f3b..d5b3aa3 100644 --- a/evals/team-tasks/vo-x8/steps/compliance/instruction.md +++ b/evals/team-tasks/vo-x8/steps/compliance/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-compliance skill. You are the `compliance` reviewer. -Review the change between git refs d5d5f00f761c3ba95294d72aae04380d7b91ba80 and cdac7cc610b5f82199f8d6773921842f92c674b6 in the repository in /workspace/repo. +Review the change between git refs ea652265cd8db98f0a52d601325af6de837357b9 and 5794f727be0b1a2474ac9659058a5efae840a94c in the repository in /workspace/repo. Write your findings to /workspace/review/findings.compliance.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x8/steps/correctness/instruction.md b/evals/team-tasks/vo-x8/steps/correctness/instruction.md index 2f9968e..c29271d 100644 --- a/evals/team-tasks/vo-x8/steps/correctness/instruction.md +++ b/evals/team-tasks/vo-x8/steps/correctness/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-correctness skill. You are the `correctness` reviewer. -Review the change between git refs d5d5f00f761c3ba95294d72aae04380d7b91ba80 and cdac7cc610b5f82199f8d6773921842f92c674b6 in the repository in /workspace/repo. +Review the change between git refs ea652265cd8db98f0a52d601325af6de837357b9 and 5794f727be0b1a2474ac9659058a5efae840a94c in the repository in /workspace/repo. Write your findings to /workspace/review/findings.correctness.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x8/steps/language/instruction.md b/evals/team-tasks/vo-x8/steps/language/instruction.md index 3de2722..461faa6 100644 --- a/evals/team-tasks/vo-x8/steps/language/instruction.md +++ b/evals/team-tasks/vo-x8/steps/language/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-language skill. You are the `language` reviewer. -Review the change between git refs d5d5f00f761c3ba95294d72aae04380d7b91ba80 and cdac7cc610b5f82199f8d6773921842f92c674b6 in the repository in /workspace/repo. +Review the change between git refs ea652265cd8db98f0a52d601325af6de837357b9 and 5794f727be0b1a2474ac9659058a5efae840a94c in the repository in /workspace/repo. Write your findings to /workspace/review/findings.language.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x8/steps/security/instruction.md b/evals/team-tasks/vo-x8/steps/security/instruction.md index 201c3f1..cac6113 100644 --- a/evals/team-tasks/vo-x8/steps/security/instruction.md +++ b/evals/team-tasks/vo-x8/steps/security/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-security skill. You are the `security` reviewer. -Review the change between git refs d5d5f00f761c3ba95294d72aae04380d7b91ba80 and cdac7cc610b5f82199f8d6773921842f92c674b6 in the repository in /workspace/repo. +Review the change between git refs ea652265cd8db98f0a52d601325af6de837357b9 and 5794f727be0b1a2474ac9659058a5efae840a94c in the repository in /workspace/repo. Write your findings to /workspace/review/findings.security.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x8/steps/tests/instruction.md b/evals/team-tasks/vo-x8/steps/tests/instruction.md index e7a77d0..d325b95 100644 --- a/evals/team-tasks/vo-x8/steps/tests/instruction.md +++ b/evals/team-tasks/vo-x8/steps/tests/instruction.md @@ -1,6 +1,6 @@ The skills are installed in: /opt/review-skills Use the review-tests skill. You are the `tests` reviewer. -Review the change between git refs d5d5f00f761c3ba95294d72aae04380d7b91ba80 and cdac7cc610b5f82199f8d6773921842f92c674b6 in the repository in /workspace/repo. +Review the change between git refs ea652265cd8db98f0a52d601325af6de837357b9 and 5794f727be0b1a2474ac9659058a5efae840a94c in the repository in /workspace/repo. Write your findings to /workspace/review/findings.tests.json. Write no other file and do not edit the repository. Before anything else, print your skill file on its own, in a command that prints nothing else, and read it to its last line. It is long. If the output says it was truncated or stops early, print the missing lines by line range until you have read all of it. diff --git a/evals/team-tasks/vo-x8/steps/validator/instruction.md b/evals/team-tasks/vo-x8/steps/validator/instruction.md index 7feb71d..8209f53 100644 --- a/evals/team-tasks/vo-x8/steps/validator/instruction.md +++ b/evals/team-tasks/vo-x8/steps/validator/instruction.md @@ -1,2 +1,2 @@ The skills are installed in: /opt/review-skills -Use the review-validator skill. The output folder is /workspace/review. The base ref is d5d5f00f761c3ba95294d72aae04380d7b91ba80. +Use the review-validator skill. The output folder is /workspace/review. The base ref is ea652265cd8db98f0a52d601325af6de837357b9. diff --git a/evals/team-tasks/vo-x8/steps/validator/workdir/setup.sh b/evals/team-tasks/vo-x8/steps/validator/workdir/setup.sh index 762fbf9..2996fcc 100755 --- a/evals/team-tasks/vo-x8/steps/validator/workdir/setup.sh +++ b/evals/team-tasks/vo-x8/steps/validator/workdir/setup.sh @@ -5,4 +5,4 @@ set -euo pipefail rm -f -- "$0" cd /workspace/repo -python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base d5d5f00f761c3ba95294d72aae04380d7b91ba80 --head cdac7cc610b5f82199f8d6773921842f92c674b6 +python3 /opt/review-skills/review-orchestrator/scripts/aggregate.py first-pass --out /workspace/review --repo . --base ea652265cd8db98f0a52d601325af6de837357b9 --head 5794f727be0b1a2474ac9659058a5efae840a94c