Skip to content

Refresh stale default Antigravity and repair model chains (#2023) - #2046

Merged
wwind123 merged 2 commits into
mainfrom
agent-loop/managed-2023
Nov 20, 2026
Merged

wwind123 merged 2 commits into
mainfrom
agent-loop/managed-2023

Conversation

@wwind123

Copy link
Copy Markdown
Owner

Summary

This refreshes the stale built-in model chains, as the approved plan asks.

  • DEFAULT_ANTIGRAVITY_MODELS is now Gemini 3.8 Flash (High) → Claude Opus 5.5 (Medium). Opus runs on a separate quota and stays last.
  • DEFAULT_REPAIR_MODELS is now Gemini 3.8 Flash (Medium) → Claude Opus 5.5 (Medium). It is still a two-model tuple.
  • The repair-model guard (if self.repair_models == DEFAULT_REPAIR_MODELS) is unchanged. So are the explicit-flag paths and review_images_capabilities.py.
  • The quota groups need no change: the prefix rule already maps the new entries to gemini and then claude, and a new test pins this.
  • With these defaults, Opus appears in both chains. The default repair walk therefore de-duplicates to [Gemini 3.8 Flash (Medium), Claude Opus 5.5 (Medium), Gemini 3.8 Flash (High)], and a test asserts that order.

Tests

  • The pins for both tuples and for the CLI help text now also assert that none of the removed names, and not Gemini 4 Hydrogen, appear in either chain.
  • New walk tests on the default chain:
    • Gemini quota exhaustion jumps straight to Opus.
    • Transient failures end at Opus with the ordinary error and no quota memory.
  • Walk-mechanics tests that need several Gemini entries now pass a test-local multi-Gemini chain.
  • New test: --repair-backend codex without --repair-model is rejected, and with --repair-model M it yields ("M",).
  • The docs-guidance test checks that both default chains appear in README.md and docs/local_agent_loop.md.
  • I ran the focused files and the full suite through agent-loop run-tests. Both passed.

Fixes #2023

AGENT_MANAGED_CI_UNPROTECTED_OVERRIDE_V1 nonce=5vXFxMMMT0mabgO8kdITug7K

🤖 Generated with Claude Code

DEFAULT_ANTIGRAVITY_MODELS becomes Gemini 3.8 Flash (High) -> Claude Opus
5.5 (Medium), and DEFAULT_REPAIR_MODELS becomes Gemini 3.8 Flash (Medium)
-> Claude Opus 5.5 (Medium). The removed Gemini 3.7/3.6 Flash and 3.1 Pro
names are gone from both chains. The repair-model guard is unchanged.

Pin tests, default-chain walk tests and docs are updated. Walk-mechanics
tests that need several gemini-group entries use a test-local chain.

Agent-Issue-Provenance: v1 repo=wwind123/coding-review-agent-loop issue=2023 flow=approved plan=02c49f6970b84553

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@wwind123 wwind123 added the agent-loop-managed agent-loop managed CI label Nov 20, 2026
@wwind123

Copy link
Copy Markdown
Owner Author

Agent-loop managed-CI authorization record (machine-readable). Binds issue #2023 to pull request #2046 at head dfcc84f. Not an agent response; keep this comment.

@wwind123

Copy link
Copy Markdown
Owner Author

Expected closing issues for PR #2046: #2023.

Origin flow: approved-plan-implementation
Contract hash: 08a12468c78c86634ed12957477419c60b7c19ccc1741eeaf7b5815b6374c6d3

-- coding-review-agent-loop

@wwind123

Copy link
Copy Markdown
Owner Author

Issue implementation

I implemented the approved plan for #2023 and opened draft PR #2046 (managed label). DEFAULT_ANTIGRAVITY_MODELS is now (Gemini 3.8 Flash (High), Claude Opus 5.5 (Medium)). DEFAULT_REPAIR_MODELS is now (Gemini 3.8 Flash (Medium), Claude Opus 5.5 (Medium)). The repair-model guard, the explicit-flag paths and review_images_capabilities.py are untouched. README.md and docs/local_agent_loop.md now name the new chains. Tests: the pins assert both tuples, the help text, and that no removed name or Gemini 4 Hydrogen is present. New pins cover the gemini/claude quota-group derivation and two default-chain walk tests (quota exhaustion jumps to Opus; transient failures end at Opus). The repair-walk test is renamed and asserts the de-duplicated three-model order. A new test covers codex repair without and with --repair-model. Multi-Gemini walk-mechanics tests now use a test-local chain. Test selection: the focused files cover each changed test module and config.py. The full suite was run because the approved plan's acceptance criteria explicitly require it, and it passed.

Result

Pull request reported: #2046.

Tests run

  • /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py (agent-loop instrumented; whole-command timeout 1800s)
  • /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/ (agent-loop instrumented; whole-command timeout 1800s)

Test observation receipts

  • '<path-sha256:d094488e6a9166dc>' -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py — receipt b840852fdd224015842f520943783038 — current-result (verified against the parent journal; CI remains authoritative)
  • '<path-sha256:d094488e6a9166dc>' -m pytest -q -p no:cacheprovider tests/ — receipt a9a0bab1fad3449cb17f851445054441 — current-result (verified against the parent journal; CI remains authoritative)

Risk-based mode and transition test matrix evidence

  • Matrix identity: a08037df81e043cd4d2cc1407f3eb5f0aea643de1daa797e1e1b47848b5d7290
Full matrix evidence (4 rows)
  • default-chain-quota-fallback — verified
    • Tests: tests/test_antigravity_quota_groups.py::test_default_chain_quota_exhaustion_jumps_straight_to_opus, tests/test_antigravity_quota_groups.py::test_default_chain_quota_groups_are_gemini_then_claude
    • Locations: tests/test_antigravity_quota_groups.py
    • Workflow path: Antigravity validated turn (_run_validated_agent) with the default chain and empty run quota memory
    • Expected outcome assertions: A verified quota frame on Gemini 3.8 Flash (High) jumps to Claude Opus 5.5 (Medium), which succeeds; the invoked models are exactly [Gemini 3.8 Flash (High), Claude Opus 5.5 (Medium)].; The default entries derive to quota groups gemini then claude.
    • Forbidden-effect assertions: No sleep command is issued and the exhausted model is not retried.; No removed Gemini model is invoked.
    • Evidence citations: /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py [b840852fdd224015842f520943783038; current-result]
  • default-chain-transient-end — verified
    • Tests: tests/test_antigravity_quota_groups.py::test_default_chain_transient_failures_end_at_opus_with_ordinary_error
    • Locations: tests/test_antigravity_quota_groups.py
    • Workflow path: Antigravity validated turn with the default chain and agent_max_retries=1, all attempts high-traffic
    • Expected outcome assertions: The invoked order is exactly [Gemini 3.8 Flash (High), Gemini 3.8 Flash (High), Claude Opus 5.5 (Medium)].; The turn ends with a plain AgentInvocationError and no gemini/claude quota memory entry.
    • Forbidden-effect assertions: The error is not a QuotaResetExceededError and does not start with 'Antigravity unavailable'.
    • Evidence citations: /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py [b840852fdd224015842f520943783038; current-result]
  • default-repair-walk — verified
    • Tests: tests/test_repair.py::test_execute_repair_default_chain_is_repair_then_deduplicated_antigravity_chain
    • Locations: tests/test_repair.py
    • Workflow path: execute_repair with the default antigravity repair backend where every repaired output fails validation
    • Expected outcome assertions: The distinct models tried are exactly [Gemini 3.8 Flash (Medium), Claude Opus 5.5 (Medium), Gemini 3.8 Flash (High)].
    • Forbidden-effect assertions: Run quota memory is unchanged after repair.; No removed model appears in the tried list.
    • Evidence citations: /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py [b840852fdd224015842f520943783038; current-result]
  • explicit-flags-unchanged — verified
    • Tests: tests/test_cli_repair.py::test_codex_repair_flag_without_model_rejected_and_single_model_verbatim, tests/test_cli_repair.py::test_cli_repair_requires_explicit_model, tests/test_cli_repair.py::test_repair_options_reach_config, tests/test_agent_loop.py
    • Locations: tests/test_cli_repair.py, tests/test_antigravity.py, tests/test_agent_loop.py
    • Workflow path: config_from_args / AgentLoopConfig validation with explicit --repair-backend codex and --repair-model, and with explicit --antigravity-model(s)
    • Expected outcome assertions: Codex repair without --repair-model raises the explicit --repair-model error.; --repair-model M yields repair_models == ('M',), and repeated models are kept verbatim.; The existing explicit --antigravity-model tests pass unmodified.
    • Forbidden-effect assertions: Defaults are not merged into an explicit repair chain.; The repair-model guard line in config.py is not changed by the diff.
    • Evidence citations: /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py [b840852fdd224015842f520943783038; current-result]

Risk-matrix coverage map

Orchestrator completeness check (existence and level only; whether each cited test really exercises its row is for the reviewer to judge).

  • default-chain-quota-fallback: tests tests/test_antigravity_quota_groups.py::test_default_chain_quota_exhaustion_jumps_straight_to_opus, tests/test_antigravity_quota_groups.py::test_default_chain_quota_groups_are_gemini_then_claude, tests/test_antigravity_quota_groups.py; declared level unit vs required unit
  • default-chain-transient-end: tests tests/test_antigravity_quota_groups.py::test_default_chain_transient_failures_end_at_opus_with_ordinary_error, tests/test_antigravity_quota_groups.py; declared level unit vs required unit
  • default-repair-walk: tests tests/test_repair.py::test_execute_repair_default_chain_is_repair_then_deduplicated_antigravity_chain, tests/test_repair.py; declared level unit vs required unit
  • explicit-flags-unchanged: tests tests/test_cli_repair.py::test_codex_repair_flag_without_model_rejected_and_single_model_verbatim, tests/test_cli_repair.py::test_cli_repair_requires_explicit_model, tests/test_cli_repair.py::test_repair_options_reach_config, tests/test_agent_loop.py, tests/test_cli_repair.py, tests/test_antigravity.py, tests/test_agent_loop.py; declared level unit vs required unit
  • Status: complete at dfcc84f (deterministic completeness check passed)

-- Anthropic Claude: claude-opus-5-5 (medium)

@wwind123

Copy link
Copy Markdown
Owner Author

Agent-loop managed-CI resume provenance audit record (machine-readable). Records resume provenance for managed CI on pull request #2046 at head dfcc84f. Not an agent response; keep this comment.

AGENT_MANAGED_CI_UNPROTECTED_OVERRIDE_V1 nonce=OSgBks5NbLKkRkDxtdE8S5BYUNJiy7ha repo=wwind123/coding-review-agent-loop base=main head=dfcc84fd562af819bb15126fb8a0c4c726bf0b40 protection=voluntary active_label_event_id=36787978984 resume_from=7044401610 provenance_head=dfcc84fd562af819bb15126fb8a0c4c726bf0b40 generation=CKqk1GWByEMCS1cx

Resume provenance only: the prior issue-created audit is not an authorization token.

@wwind123

Copy link
Copy Markdown
Owner Author

PR review scheduling audit: selected Codex; paused Claude, Antigravity; reason: strict pre-panel fallback: architecture identity changed; scheduler metadata recovery: latest scheduler metadata is missing or invalid; scheduler metadata recovery: current-head records lack scheduler metadata; scheduler metadata is missing or invalid; primary re-invoked with full context; phase: primary; head: dfcc84f; primary: Codex; active owners: (none); force-full: False (source: none); scheduler-policy calls avoided cumulatively: 2.

@wwind123

Copy link
Copy Markdown
Owner Author

Review verdict: Blocking

Reviewed the complete diff at dfcc84f. The requested constants, documentation, explicit-flag behavior, and four risk-matrix rows are supported by source inspection and supplied passing receipts; GitHub checks are passing. One existing chain-walk test loses its intended coverage. No tests were run during review.

Blocking issues

  • [P2] Preserve the transcript-reset end-of-chain test coverage (tests/test_antigravity_quota_groups.py:330-340). With the new two-model default and agent_max_retries=1, this turn invokes Gemini twice and Opus once, then stops. Its special noise + HIGH_TRAFFIC output is still queued sixth, after five ordinary HIGH_TRAFFIC failures, so it is never consumed. The test now passes without exercising an unverified reset at the end of the chain, contrary to the approved requirement to preserve walk-mechanics coverage. Queue the special output third for the new default, or pass MULTI_GEMINI_CHAIN and adjust the expected sequence; assert that the special output was actually consumed.

-- OpenAI Codex: gpt-6.1-sol (medium)

@wwind123

Copy link
Copy Markdown
Owner Author

PR review round 1 reconciliation: settled reviewers: Codex. Finalization continues after reconciliation. Historical approvals remain exact-head-bound; scheduler-policy calls avoided cumulatively: 2. Phase: primary; force-full: False (source: none).

@wwind123

Copy link
Copy Markdown
Owner Author

PR #2046 coder dispatch record: round 1, attempt 1, head dfcc84fd562af819bb15126fb8a0c4c726bf0b40. Written before the coder is invoked so an interrupted or rejected turn can be resumed.

The transcript-reset test queued six outputs but the two-entry default
chain consumes three, so its noise output was never read. Parametrize it
over the test-local multi-Gemini chain and the default chain, queue one
output per expected call, and assert every output was consumed. The new
default-chain transient test gets the same exact-queue assertion.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@wwind123

Copy link
Copy Markdown
Owner Author

Coder follow-up

Fixed item-1 by restoring end-of-chain transcript-reset coverage, pushed as commit 282852e. The test is now parametrized over the test-local multi-Gemini chain (the original five-call walk) and the new default chain (three calls). It queues exactly one output per expected call and asserts that the queue is empty afterwards, so the noise output on Opus is provably consumed. Generalization: this generalizes the fix for [item-1]: a walk test must queue exactly the outputs its expected walk consumes and assert that all of them were consumed. I applied the same rule to the new default-chain transient test, which also over-queued (six outputs for three calls). Test rationale: tests/test_antigravity_quota_groups.py is the only changed file and holds every test touched here.

Addressed items

  • item-1: Blocking issue from OpenAI Codex: gpt-6.1-sol (medium), round 1: [P2] Preserve the transcript-reset end-of-chain test coverage (tests/test_antigravity_quota_group...
    • Resolution: test_unverified_transcript_reset_at_end_of_chain_does_not_raise_quota_reset is now parametrized over [multi-gemini, default]. It passes an explicit runner and chain, queues len(expected_models) outputs with the noise output last, and asserts models == [GEM1, GEM1, *chain[1:]] and runner.antigravity_outputs == []. The default-chain transient test now queues exactly three failures and asserts that all of them were consumed. Run: tests/test_antigravity_quota_groups.py (84 passed).

Remaining items

  • None.

Tests run

  • /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider -n 0 tests/test_antigravity_quota_groups.py (agent-loop instrumented; whole-command timeout 1800s)

Test observation receipts

  • '<path-sha256:d094488e6a9166dc>' -m pytest -q -p no:cacheprovider -n 0 tests/test_antigravity_quota_groups.py — receipt 1060beda6cf84e33a552a84481b47012 — current-result (verified against the parent journal; CI remains authoritative)

Risk-based mode and transition test matrix evidence

  • Matrix identity: a08037df81e043cd4d2cc1407f3eb5f0aea643de1daa797e1e1b47848b5d7290
4 changed rows; 0 unchanged since round 1
  • default-repair-walk — missing
    • Tests: none
    • Locations: none
    • Workflow path: execute_repair with default antigravity repair backend
    • Expected outcome assertions: The distinct models tried are exactly [Gemini 3.8 Flash (Medium), Claude Opus 5.5 (Medium), Gemini 3.8 Flash (High)].
    • Forbidden-effect assertions: Do not read or mark run quota memory.; Do not try any removed model.
    • Caveats: No semantic coverage claim was supplied for this enforceable row.
  • explicit-flags-unchanged — missing
    • Tests: none
    • Locations: none
    • Workflow path: config_from_args / AgentLoopConfig validation
    • Expected outcome assertions: Explicit chains are used verbatim. Codex repair without --repair-model is rejected, and with --repair-model M it yields ("M",).
    • Forbidden-effect assertions: Do not merge defaults into an explicit chain.; Do not change the repair-model guard.
    • Caveats: No semantic coverage claim was supplied for this enforceable row.
  • default-chain-quota-fallback — verified
    • Tests: tests/test_antigravity_quota_groups.py::test_default_chain_quota_exhaustion_jumps_straight_to_opus, tests/test_antigravity_quota_groups.py::test_default_chain_quota_groups_are_gemini_then_claude
    • Locations: tests/test_antigravity_quota_groups.py
    • Workflow path: Antigravity validated turn (_run_validated_agent) with the default chain and empty run quota memory
    • Expected outcome assertions: A verified quota frame on Gemini 3.8 Flash (High) jumps to Claude Opus 5.5 (Medium), which succeeds; the invoked models are exactly [Gemini 3.8 Flash (High), Claude Opus 5.5 (Medium)].
    • Forbidden-effect assertions: No sleep command is issued and the exhausted model is not retried; no removed Gemini model is invoked.
    • Evidence citations: /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider -n 0 tests/test_antigravity_quota_groups.py [1060beda6cf84e33a552a84481b47012; current-result]
  • default-chain-transient-end — verified
    • Tests: tests/test_antigravity_quota_groups.py::test_default_chain_transient_failures_end_at_opus_with_ordinary_error, tests/test_antigravity_quota_groups.py::test_unverified_transcript_reset_at_end_of_chain_does_not_raise_quota_reset[default]
    • Locations: tests/test_antigravity_quota_groups.py
    • Workflow path: Antigravity validated turn with the default chain and agent_max_retries=1, all attempts high-traffic
    • Expected outcome assertions: The invoked order is exactly [Gemini 3.8 Flash (High), Gemini 3.8 Flash (High), Claude Opus 5.5 (Medium)], every queued output is consumed, and the turn ends with a plain AgentInvocationError with no quota memory recorded.
    • Forbidden-effect assertions: The error is not a QuotaResetExceededError and does not start with 'Antigravity unavailable'.
    • Evidence citations: /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider -n 0 tests/test_antigravity_quota_groups.py [1060beda6cf84e33a552a84481b47012; current-result]

0 rows unchanged since round 1; full matrix in round 1.

Risk-matrix coverage map

Orchestrator completeness check (existence and level only; whether each cited test really exercises its row is for the reviewer to judge).

  • default-chain-quota-fallback: tests tests/test_antigravity_quota_groups.py::test_default_chain_quota_exhaustion_jumps_straight_to_opus, tests/test_antigravity_quota_groups.py::test_default_chain_quota_groups_are_gemini_then_claude, tests/test_antigravity_quota_groups.py; declared level unit vs required unit
  • default-chain-transient-end: tests tests/test_antigravity_quota_groups.py::test_default_chain_transient_failures_end_at_opus_with_ordinary_error, tests/test_antigravity_quota_groups.py::test_unverified_transcript_reset_at_end_of_chain_does_not_raise_quota_reset[default], tests/test_antigravity_quota_groups.py; declared level unit vs required unit
  • default-repair-walk: no test claim and no declared gap (required unit)
  • explicit-flags-unchanged: no test claim and no declared gap (required unit)
  • Status: incomplete at 282852e: default-repair-walk: missing-row; explicit-flags-unchanged: missing-row

-- Anthropic Claude: claude-opus-5-5 (medium)

@wwind123

Copy link
Copy Markdown
Owner Author

Agent-loop managed-CI authorization continuity record (machine-readable). Binds issue #2023 to pull request #2046 at head 282852e. Not an agent response; keep this comment.

@wwind123

Copy link
Copy Markdown
Owner Author

PR review scheduling audit: selected Codex; paused Claude, Antigravity; reason: primary phase: primary rechecks its findings before the secondary panel; phase: primary; head: 282852e; primary: Codex; active owners: Codex; force-full: False (source: none); scheduler-policy calls avoided cumulatively: 4.

@wwind123

Copy link
Copy Markdown
Owner Author

Review verdict: Blocking

Reviewed the complete base-to-head diff at 282852e. Item-1 is fixed and no further code defects were found. GitHub checks are passing, but two enforceable risk-matrix rows lack admissible coverage citations. No tests were run during review.

Blocking issues

  • [P2] Supply passing captured evidence for default-repair-walk. The coverage map reports this row missing. tests/test_repair.py::test_execute_repair_default_chain_is_repair_then_deduplicated_antigravity_chain exercises execute_repair and asserts the required three-model order and unchanged quota memory, but the supplied broader receipts are marked stale and the current-head receipt covers only test_antigravity_quota_groups.py. Run the existing repair-walk test through the capture runner on the reviewed tree and cite its passing observation in the row's semantic coverage claim.
  • [P2] Supply passing captured evidence for explicit-flags-unchanged. The coverage map reports this row missing. Existing tests cover explicit Antigravity chains, the single-model flag, and Codex repair rejection/explicit-model precedence, but none has a supplied current-tree passing citation. Capture passing runs of the relevant tests in test_antigravity.py, test_agent_loop.py and test_cli_repair.py, including test_codex_repair_flag_without_model_rejected_and_single_model_verbatim, and attach their observations to this row's semantic coverage claim. The older broader receipts are explicitly stale; the current-head quota-group receipt does not cover these tests.

Prior unresolved item dispositions

  • [item-1] RESOLVED: The transcript-reset test now covers both chains, queues the noisy output on the final Opus invocation, asserts the complete expected model sequence, and verifies the output queue is empty. The supplied current-head quota-group receipt passes.
    • Original finding: Blocking issue from OpenAI Codex: gpt-6.1-sol (medium), round 1: [P2] Preserve the transcript-reset end-of-chain test coverage (tests/test_antigravity_quota_group...

-- OpenAI Codex: gpt-6.1-sol (medium)

@wwind123

Copy link
Copy Markdown
Owner Author

PR review round 2 reconciliation: settled reviewers: Codex. Finalization continues after reconciliation. Historical approvals remain exact-head-bound; scheduler-policy calls avoided cumulatively: 4. Phase: primary; force-full: False (source: none).

@wwind123

Copy link
Copy Markdown
Owner Author

PR #2046 coder dispatch record: round 2, attempt 1, head 282852ead0543db26ae5db3c890051234c473a91. Written before the coder is invoked so an interrupted or rejected turn can be resumed.

@wwind123

Copy link
Copy Markdown
Owner Author

Coder follow-up

Orchestrator note: the PR head was unchanged at 282852ead0543db26ae5db3c890051234c473a91 after this follow-up; no new commit is visible on the PR branch. The coder statements below are claims about the existing head, not changes attributed to this turn.

Coder summary (claim, unverified): Items 2 and 3 asked for evidence, not code changes, so I made no code change. I re-ran all six focused test files through agent-loop run-tests on the reviewed head 282852e (1068 passed) and cite that one current-tree run for all four risk-matrix rows. Generalization: this generalizes the fix for [item-2]: every risk-matrix row cites a passing observation from a run on the current tree, not a run on an earlier tree. Test rationale: test_cli_repair.py (repair pins and the explicit codex flag), test_antigravity.py (the default chain and explicit Antigravity chains), test_antigravity_quota_groups.py (default-chain walks), test_repair.py (default repair walk), test_docs_guidance.py (docs chain assertions) and test_agent_loop.py (explicit --antigravity-model). Each maps to a file the PR changed or to a matrix row.

Claimed already present at 282852ead0543db26ae5db3c890051234c473a91 (PR head unchanged)

  • item-2: Blocking issue from OpenAI Codex: gpt-6.1-sol (medium), round 2: [P2] Supply passing captured evidence for default-repair-walk. The coverage map reports this row...
    • Coder claim: On head 282852e, tests/test_repair.py::test_execute_repair_default_chain_is_repair_then_deduplicated_antigravity_chain passed in the captured focused run. It is cited in the default-repair-walk claim.
  • item-3: Blocking issue from OpenAI Codex: gpt-6.1-sol (medium), round 2: [P2] Supply passing captured evidence for explicit-flags-unchanged. The coverage map reports this...
    • Coder claim: On head 282852e, the explicit-flag tests in test_cli_repair.py (including test_codex_repair_flag_without_model_rejected_and_single_model_verbatim), test_antigravity.py and test_agent_loop.py passed in the same captured run. They are cited in the explicit-flags-unchanged claim.

Remaining items

  • None.

Tests run

  • /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py (agent-loop instrumented; whole-command timeout 1800s)

Test observation receipts

  • '<path-sha256:d094488e6a9166dc>' -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py — receipt cfbbe80b59bf42c0801139f33d66c7e0 — current-result (verified against the parent journal; CI remains authoritative)

Risk-based mode and transition test matrix evidence

  • Matrix identity: a08037df81e043cd4d2cc1407f3eb5f0aea643de1daa797e1e1b47848b5d7290
4 changed rows; 0 unchanged since round 2
  • default-chain-quota-fallback — verified
    • Tests: tests/test_antigravity_quota_groups.py::test_default_chain_quota_exhaustion_jumps_straight_to_opus, tests/test_antigravity_quota_groups.py::test_default_chain_quota_groups_are_gemini_then_claude
    • Locations: tests/test_antigravity_quota_groups.py
    • Workflow path: Antigravity validated turn (_run_validated_agent) with the default chain and empty run quota memory
    • Expected outcome assertions: A verified quota frame on Gemini 3.8 Flash (High) jumps to Claude Opus 5.5 (Medium), which succeeds; the invoked models are exactly [Gemini 3.8 Flash (High), Claude Opus 5.5 (Medium)].
    • Forbidden-effect assertions: No sleep command is issued, the exhausted model is not retried, and no removed Gemini model is invoked.
    • Evidence citations: /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py [cfbbe80b59bf42c0801139f33d66c7e0; current-result]
  • default-chain-transient-end — verified
    • Tests: tests/test_antigravity_quota_groups.py::test_default_chain_transient_failures_end_at_opus_with_ordinary_error, tests/test_antigravity_quota_groups.py::test_unverified_transcript_reset_at_end_of_chain_does_not_raise_quota_reset
    • Locations: tests/test_antigravity_quota_groups.py
    • Workflow path: Antigravity validated turn with the default chain and agent_max_retries=1, all attempts high-traffic
    • Expected outcome assertions: The invoked order is exactly [Gemini 3.8 Flash (High), Gemini 3.8 Flash (High), Claude Opus 5.5 (Medium)], every queued output is consumed, and the turn ends with a plain AgentInvocationError with no quota memory recorded.
    • Forbidden-effect assertions: The error is not a QuotaResetExceededError and does not start with 'Antigravity unavailable'.
    • Evidence citations: /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py [cfbbe80b59bf42c0801139f33d66c7e0; current-result]
  • default-repair-walk — verified
    • Tests: tests/test_repair.py::test_execute_repair_default_chain_is_repair_then_deduplicated_antigravity_chain
    • Locations: tests/test_repair.py
    • Workflow path: execute_repair with the default antigravity repair backend where every repaired output fails validation
    • Expected outcome assertions: The distinct models tried are exactly [Gemini 3.8 Flash (Medium), Claude Opus 5.5 (Medium), Gemini 3.8 Flash (High)].
    • Forbidden-effect assertions: Run quota memory is unchanged after repair, and no removed model appears in the tried list.
    • Evidence citations: /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py [cfbbe80b59bf42c0801139f33d66c7e0; current-result]
  • explicit-flags-unchanged — verified
    • Tests: tests/test_cli_repair.py::test_codex_repair_flag_without_model_rejected_and_single_model_verbatim, tests/test_cli_repair.py::test_cli_repair_requires_explicit_model, tests/test_cli_repair.py::test_repair_options_reach_config
    • Locations: tests/test_cli_repair.py, tests/test_antigravity.py, tests/test_agent_loop.py
    • Workflow path: config_from_args / AgentLoopConfig validation with explicit --repair-backend codex and --repair-model, and with explicit --antigravity-model(s)
    • Expected outcome assertions: Codex repair without --repair-model raises the explicit --repair-model error; --repair-model M yields ('M',); repeated models are kept verbatim; the existing explicit --antigravity-model(s) tests in test_antigravity.py and test_agent_loop.py pass unmodified.
    • Forbidden-effect assertions: Defaults are not merged into an explicit chain, and the repair-model guard line in config.py is not changed by the diff.
    • Evidence citations: /home/wwind123/tools/coding-review-agent-loop/.venv/bin/python -m pytest -q -p no:cacheprovider tests/test_cli_repair.py tests/test_antigravity.py tests/test_antigravity_quota_groups.py tests/test_repair.py tests/test_docs_guidance.py tests/test_agent_loop.py [cfbbe80b59bf42c0801139f33d66c7e0; current-result]

0 rows unchanged since round 2; full matrix in round 1.

Risk-matrix coverage map

Orchestrator completeness check (existence and level only; whether each cited test really exercises its row is for the reviewer to judge).

  • default-chain-quota-fallback: tests tests/test_antigravity_quota_groups.py::test_default_chain_quota_exhaustion_jumps_straight_to_opus, tests/test_antigravity_quota_groups.py::test_default_chain_quota_groups_are_gemini_then_claude, tests/test_antigravity_quota_groups.py; declared level unit vs required unit
  • default-chain-transient-end: tests tests/test_antigravity_quota_groups.py::test_default_chain_transient_failures_end_at_opus_with_ordinary_error, tests/test_antigravity_quota_groups.py::test_unverified_transcript_reset_at_end_of_chain_does_not_raise_quota_reset, tests/test_antigravity_quota_groups.py; declared level unit vs required unit
  • default-repair-walk: tests tests/test_repair.py::test_execute_repair_default_chain_is_repair_then_deduplicated_antigravity_chain, tests/test_repair.py; declared level unit vs required unit
  • explicit-flags-unchanged: tests tests/test_cli_repair.py::test_codex_repair_flag_without_model_rejected_and_single_model_verbatim, tests/test_cli_repair.py::test_cli_repair_requires_explicit_model, tests/test_cli_repair.py::test_repair_options_reach_config, tests/test_cli_repair.py, tests/test_antigravity.py, tests/test_agent_loop.py; declared level unit vs required unit
  • Status: complete at 282852e (deterministic completeness check passed)

-- Anthropic Claude: claude-opus-5-5 (medium)

@wwind123

Copy link
Copy Markdown
Owner Author

PR review scheduling audit: selected Codex; paused Claude, Antigravity; reason: strict pre-panel fallback: the active obligation ledger is not reconstructible; primary re-invoked with full context; phase: primary; head: 282852e; primary: Codex; active owners: Codex; force-full: False (source: none); scheduler-policy calls avoided cumulatively: 6.

@wwind123

Copy link
Copy Markdown
Owner Author

Review verdict: Approved

Reviewed the complete base-to-head diff at 282852e. The constants, documentation, default-chain walks and explicit-flag coverage satisfy the approved scope. All four matrix rows have suitable tests and a current-head parent-observed passing citation (cfbbe80b59bf42c0801139f33d66c7e0), with a reported environment-difference caveat. Both carried evidence gaps are resolved. GitHub checks are passing; no tests were run during this review.

Prior unresolved item dispositions

  • [item-2] RESOLVED: The current-head parent-observed passing receipt cfbbe80b59bf42c0801139f33d66c7e0 includes test_repair.py and is cited for default-repair-walk; source inspection confirms the required three-model order and unchanged quota-memory assertion.
    • Original finding: Blocking issue from OpenAI Codex: gpt-6.1-sol (medium), round 2: [P2] Supply passing captured evidence for default-repair-walk. The coverage map reports this row...
  • [item-3] RESOLVED: The same current-head passing receipt includes test_cli_repair.py, test_antigravity.py and test_agent_loop.py and is cited for explicit-flags-unchanged. Inspected tests cover explicit chains, the single-model flag, and Codex repair rejection and explicit-model precedence; the repair guard remains unchanged.
    • Original finding: Blocking issue from OpenAI Codex: gpt-6.1-sol (medium), round 2: [P2] Supply passing captured evidence for explicit-flags-unchanged. The coverage map reports this...

-- OpenAI Codex: gpt-6.1-sol (medium)

@wwind123

Copy link
Copy Markdown
Owner Author

PR review round 3 reconciliation: settled reviewers: Codex. Finalization continues after reconciliation. Historical approvals remain exact-head-bound; scheduler-policy calls avoided cumulatively: 6. Phase: primary; force-full: False (source: none).

@wwind123

Copy link
Copy Markdown
Owner Author

PR review scheduling audit: selected Claude, Antigravity; paused Codex; reason: independent secondary audit after exact-head primary approval; phase: secondary-audit; head: 282852e; primary: Codex; active owners: (none); force-full: False (source: none); scheduler-policy calls avoided cumulatively: 6.

@wwind123

Copy link
Copy Markdown
Owner Author

Review verdict: Approved

I reviewed the full base-to-head diff at 282852e on my own. config.py sets DEFAULT_ANTIGRAVITY_MODELS to exactly (Gemini 3.8 Flash (High), Claude Opus 5.5 (Medium)), with the separate-quota Opus entry last. DEFAULT_REPAIR_MODELS is the two-model tuple (Gemini 3.8 Flash (Medium), Claude Opus 5.5 (Medium)). The diff does not touch the repair-model guard, review_images_capabilities.py or any explicit-flag path. Quota groups still come from the name prefix, and a new pin checks that they are [gemini, claude]. All four risk-matrix rows are covered by tests that run the real path. (1) test_default_chain_quota_exhaustion_jumps_straight_to_opus checks the order [GEM1, OPUS], with no sleep and no removed models. (2) test_default_chain_transient_failures_end_at_opus_with_ordinary_error checks the order [GEM1, GEM1, OPUS], that every queued output is consumed, that no quota memory is written, and that the error is a plain AgentInvocationError. (3) The renamed repair test checks the de-duplicated order [3.8 Flash (Medium), Opus, 3.8 Flash (High)] and that quota memory is unchanged. (4) The new codex --repair-model test checks both the rejection and the verbatim ('M',). Tests that check walk mechanics now pass an explicit test-local MULTI_GEMINI_CHAIN, so they keep their four-Gemini coverage, and the end-of-chain transcript-noise test runs on both chains. README.md and docs/local_agent_loop.md name the new default and repair chains, and the docs-guidance test checks both files and rejects the removed names in the chain sentences. A grep of src/, README.md and the guide finds no default-chain use of the removed models. One left-over: the illustrative explicit-chain example at docs/local_agent_loop.md:1144 still names Gemini 3.1 Pro (High). That example is not a default, and models missing from agy models are skipped, so it does not block. GitHub checks are passing. The coder's current-head focused run (receipt cfbbe80b) passed; I did not run tests myself.

-- Anthropic Claude: claude-opus-5-5 (medium)

@wwind123

Copy link
Copy Markdown
Owner Author

Review verdict: Approved

Approved. The PR cleanly refreshes DEFAULT_ANTIGRAVITY_MODELS and DEFAULT_REPAIR_MODELS to match current provider availability, omitting stale Gemini 3.7/3.6/3.1 entries and Gemini 4 Hydrogen. The repair-model guard and explicit-flag precedence remain preserved and unmodified. Tests pin the new tuples, verify quota group derivations, test default-chain quota fallback and transient exhaustion, and preserve multi-model gemini group walk mechanics via a test fixture. Docs are accurately updated.

-- Google Antigravity: Gemini 3.8 Flash (High)

@wwind123

Copy link
Copy Markdown
Owner Author

PR review round 4 reconciliation: settled reviewers: Antigravity, Claude. Finalization continues after reconciliation. Historical approvals remain exact-head-bound; scheduler-policy calls avoided cumulatively: 6. Phase: secondary-audit; force-full: False (source: none).

@wwind123

wwind123 commented Nov 20, 2026 •

Copy link
Copy Markdown
Owner Author

Managed CI authorization for exact head 282852e.

@wwind123
wwind123 marked this pull request as ready for review November 20, 2026 16:33
@wwind123
wwind123 merged commit 8744d5c into main Nov 20, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Refresh stale default Antigravity and repair model chains

1 participant