measurement: timing, noise rule and precheck in a new package; task image builds as a runner; short prompts by default by ArjunS07 · Pull Request #66 · formula-code/datasmith

verify: fc-verify with status, classify, report, review and automatic not_approved rules

ArjunS07

committed

ef70351

verify: precheck stage (patch applies, base edits, benchmark coverage) with the in-image check

ArjunS07

committed

7fcb5eb

verify: build, pin, scrub and publish stages for task images

ArjunS07

committed

61f41a8

verify: measure stage, measure gate service and NopTouch; template logs the gate lease id

ArjunS07

committed

37fb966

verify: test reads the job config with Path

ArjunS07

committed

964e046

verify: release readiness (registry, template commits, runs, CPU types, license, PR author, prompt audit); register measure in the CLI

ArjunS07

committed

c5f37e1

verify: snapshot capture stage

ArjunS07

committed

3d3c26b

verify: port the build guard, disk stop and image check tests

ArjunS07

committed

2ed3f64

verify: PR number from the image name too

ArjunS07

committed

493750c

verify: an empty pin reads as no pin

ArjunS07

committed

6759111

datasmith: load dotenv only when tokens.env exists, so the verify stages import with the standard library

ArjunS07

committed

08dca8a

verify: the command imports with the standard library, so build hosts run it with system python

ArjunS07

committed

f76f520

verify: push verified manifest rows to the registry as they are

ArjunS07

committed

c8ace12

verify: select (records, database candidates, sampler) and env-image stages

ArjunS07

committed

a2faa02

verify: port the sampler order and replacement tests

ArjunS07

committed

764896c

Merge remote-tracking branch 'origin/release/template-v1' into verify/core

ArjunS07

committed

7887c21

verify: apply approved prompt text to rendered tasks

ArjunS07

committed

fe662ba

prompts: short (symptom-only) prompts by default for new tasks, a prompt check, model drafts for review

ArjunS07

committed

10eeee4

Revert an unrelated lint change to harbor_healthcheck.py

ArjunS07

committed

38c2048

Pin harbor to formulacode/harbor-reliability e7d19622, the commit formulacode-verified-rl uses

ArjunS07

committed

9f97aa9

Merge remote-tracking branch 'origin/release/template-v1' into verify/core

ArjunS07

committed

d10e6a6

verify: count oracle test failures from the test rerun (oracle_tests_rerun)

ArjunS07

committed

fa3ae71

prompts: rewrite handles an empty problem text

ArjunS07

committed

8f0db18

prompts: keep the blank line before the next heading when the problem text is empty (same as #67)

ArjunS07

committed

5d56364

Merge origin/release/template-v1 (template-v1.4) into verify/core

ArjunS07

committed

498c2bd

measurement: new package with the noise rule and the NopTouch agent, moved from verify

ArjunS07

committed

3964b62

harbor healthcheck: keep the oracle snapshot reference in a folder; delete verify/capture.py

ArjunS07

committed

1963ea7

measurement: the measure job moves to measurement/run.py; trials take a slot of a host lock folder instead of the HTTP measure gate

ArjunS07

committed

f80f61e

measurement: precheck decides from the image build check fields (template-v1.4); delete the in-image coverage check

ArjunS07

committed

c4dbcf6

images: task image build in docker/, the build loop as BuildTaskImagesRunner, the image pin at render time; delete verify images, pin, registry, scrub and env_image

ArjunS07

committed

2f98e4e

prompts: RenderShortProblemsRunner renders the symptom only, refuses giveaways, and is the pipeline default; drop FC_PROMPT_VERSION

ArjunS07

committed

6dce9d6

measurement: write the harbor job config as JSON (drops the yaml import); drop the unused grade hook

ArjunS07

committed

6cf5990

verify: delete the package and fc-verify; fc-measure (precheck, measure) and fc-build-tasks expose what stays

ArjunS07

committed

7f5dda7

build_task_images: look up Docker's data dir once, and fall back to / when it cannot be read

ArjunS07

committed

13b98a0

measurement, harbor_adapter: type annotations for mypy; ruff format; mark the configured Supabase URL open as allowed

ArjunS07

committed

7854274

Navigation Menu

Repository navigation

measurement: timing, noise rule and precheck in a new package; task image builds as a runner; short prompts by default - #66

measurement: timing, noise rule and precheck in a new package; task image builds as a runner; short prompts by default#66
ArjunS07 merged 35 commits into
release/template-v1from
verify/core

Commits on Oct 19, 2026

verify: fc-verify with status, classify, report, review and automatic not_approved rules

verify: precheck stage (patch applies, base edits, benchmark coverage) with the in-image check

verify: build, pin, scrub and publish stages for task images

verify: measure stage, measure gate service and NopTouch; template logs the gate lease id

verify: test reads the job config with Path

verify: release readiness (registry, template commits, runs, CPU types, license, PR author, prompt audit); register measure in the CLI

verify: snapshot capture stage

verify: port the build guard, disk stop and image check tests

verify: PR number from the image name too

verify: an empty pin reads as no pin

datasmith: load dotenv only when tokens.env exists, so the verify stages import with the standard library

verify: the command imports with the standard library, so build hosts run it with system python

verify: push verified manifest rows to the registry as they are

verify: select (records, database candidates, sampler) and env-image stages

verify: port the sampler order and replacement tests

Merge remote-tracking branch 'origin/release/template-v1' into verify/core

verify: apply approved prompt text to rendered tasks

prompts: short (symptom-only) prompts by default for new tasks, a prompt check, model drafts for review

Revert an unrelated lint change to harbor_healthcheck.py

Pin harbor to formulacode/harbor-reliability e7d19622, the commit formulacode-verified-rl uses

Merge remote-tracking branch 'origin/release/template-v1' into verify/core

verify: count oracle test failures from the test rerun (oracle_tests_rerun)

prompts: rewrite handles an empty problem text

prompts: keep the blank line before the next heading when the problem text is empty (same as #67)

Merge origin/release/template-v1 (template-v1.4) into verify/core

measurement: new package with the noise rule and the NopTouch agent, moved from verify

harbor healthcheck: keep the oracle snapshot reference in a folder; delete verify/capture.py

measurement: the measure job moves to measurement/run.py; trials take a slot of a host lock folder instead of the HTTP measure gate

measurement: precheck decides from the image build check fields (template-v1.4); delete the in-image coverage check

images: task image build in docker/, the build loop as BuildTaskImagesRunner, the image pin at render time; delete verify images, pin, registry, scrub and env_image

prompts: RenderShortProblemsRunner renders the symptom only, refuses giveaways, and is the pipeline default; drop FC_PROMPT_VERSION

measurement: write the harbor job config as JSON (drops the yaml import); drop the unused grade hook

verify: delete the package and fc-verify; fc-measure (precheck, measure) and fc-build-tasks expose what stays

build_task_images: look up Docker's data dir once, and fall back to / when it cannot be read

measurement, harbor_adapter: type annotations for mypy; ruff format; mark the configured Supabase URL open as allowed