Repository navigation
test: drop review tags and fix history from hold-area test comments - #37416
Conversation
UX Review (Opus 5.5) — ⏭️ skippedRevision |
First Principles Review (Opus 5.5) — ⏭️ skippedRevision |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing to check. The change only edits comments in test files, so no behavior, contract or surface changes. [DESIGN-REVIEWED] f23238e Description read[DESCRIPTION-READ] 8831cb0f4e95e20717eb0e7076293e14a0d15aa639cd1832ae074ccb1b2fe7be That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
|
The Backend Tests (3.12, 1/2/8), Windows (8) and Coverage Gate reds on this head are main's, not this PR's: the same row-scrub tests fail on clean main 99dc8da ( |
Code change approver recommendation: ApproveWhy: the decider estimates a 5% chance that this change introduces a defect, within the 12% limit for approving on its own, and every check passes. Comment-only change across 29 test files. It removes review tags and fix-history wording and keeps every security explanation, now written as a rule. No code, assertion, fixture or source changes, so it can't change test behavior. The failing CI checks are redaction assertions in files this PR doesn't touch and can't have been caused by comment edits. An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit |
Remove review-round, finding and severity tags and before/after fix stories from comments in the sandbox, security, governance, name-grant, platform, skill and validation test files, keeping every sentence that explains a security reason. Comments only; no test behaviour changes.
The comment above the pinning line said two values while the line pins three (id, expires_at, phase). The phase is a default state, not minted, so the comment now names the two minted values and the phase. Comment only.
02453d8 to
f23238e
Compare
KiroCrewCaptain
left a comment
There was a problem hiding this comment.
Approved. On this head, a Python AST comparison of all 30 touched files against the merge-base found no differences. Only # comments change, and no code, docstrings, assertions or noqa/flake-ok markers do. Every path is under test/. No commit on main has touched these files since the merge-base, and the sibling passes already merged share no file with this PR. The body's file and line counts match the diff.
Done well: each rewrite in test_name_grant.py keeps the attack and the rule in the present tense and drops only the review round, so the security reason now comes first in each comment.
Optional:
test/test_skill_providers.py:83: "(not an AttributeError)" reads as if.get()on a scalar does not raise. Something like "(which must not surface as an AttributeError)" says what is meant.test/test_memory_smoke.py:247: "(a fix that only a restart heals is a bug)" is unclear. Consider "(a stale cache that only a restart clears is a bug)".- Some tags and fix-history notes are still in the files this PR edits, for example
# delivered 2attest/test_computer_use_drag_path.py:312and theF1-2/F2-3tags in the governance tests. A later pass could remove them.
Problem / Motivation
Goal: Remove review tags and before/after fix stories from comments in the sandbox, security, governance, name-grant, platform, skill and validation test files, keeping every sentence that explains a security reason, with no change to test behavior.
These tests carry the densest review history in the suite:
test_name_grant.pyalone opens 28 comments with "GPT 5.6 round-N:" or "Opus 4.8 round-18.", the governance tests open with "HIGH (GPT pass 2):" and "Regression (PR #14 alice):", and many comments say "Before the fix this returned ()" instead of the rule.Why it matters
These tests are the proof behind the agent's own fences. The attack each one names is the useful part; the review round that found it is not, and it pushes the attack further down the comment.
Not a goal
test/conftest.py, and no new comments that restate the code.test_security_conductor*) or third-party product names (test_connections_registry.py), and two dense history blocks left as written (test_security_gate_liveness.pycount note,test_security_posture.pyalternation note).What changed (motivation → approach → change)
shlexdiscards the rest of the line after#..."shlexdiscards the rest of the line after#; bash does NOT ..."Twenty-nine files: 105 insertions, 108 deletions.
goal_scan.py astcheckagainst the merge-base:files checked: 29 code-changed: 0 docstring-only: 0.A second commit corrects one comment in
test/test_browser_quick_ask_contract.py: it said the test pins "the two values a real ask mints", but the line sets three fields (id,expires_at,phase). With it: thirty files, code-changed 0.Backwards compatibility
Compatible: test-only change, comments only.
Removes nothing: no code, key, path, record or function is removed or renamed.
Tests
Proof: not provable -- comment-only change; the AST check shows no code difference
The twenty-nine touched files plus
test/test_flake_pattern_ratchet.pypass on this branch (pytest -n 2 -p no:randomly: 2284 passed, 24 skipped).flake8,check_comment_history.pyandcheck_brand_name.pypass; no added line is over 100 columns.Manual verification
N/A — comment-only change.
Related Issues
N/A — part of the test comment quality pass.
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
Contributed under the project's Apache-2.0 license.