docs(project-harness): attribute verdict source and guard green runs - #121
Merged
phodal merged 1 commit intoAug 27, 2026
Merged
Conversation
agent-verify-loop.md covered whether verification exists and how strong it is, but not what an agent should do with an actual verdict. Hosted-CI red verdicts went straight into a repair loop that assumes local reproducibility, and green runs had no integrity guards beyond the zero-test case. Add a verdict-source attribution gate (local change, infrastructure, upstream asset drift) before the diagnose-repair sequence, so only failures the change owns consume repair rounds. Add green-verdict guards (empty run fails, fail fast over silent fallback, no silent exits) and two matching anti-patterns. Zero-test guidance stays in verification-environment.md and is cross-referenced, not duplicated. The new cross-reference link bumps the frozen Markdown-closure edge count in antigravity-plugin-artifact.test.mjs from 307 to 308. Validated: npm test - 104 files, 1545 passed, 2 pre-existing skips; doc-link-graph.test.mjs 8/8.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
agent-verify-loop.md: red verdicts are classified as local change / infrastructure / upstream asset drift before the diagnose-repair sequence, so only failures the change owns consume repair rounds.Why
Traceability and Scope
verification-environment.md; the change is focused and host-agnostic, covering attribution and silent fallback.references/project-harness/agent-verify-loop.md(verdict domain and post-verdict sequence already owned here; Ownership Boundary section updated to name the two additions);test/plugins/antigravity-plugin-artifact.test.mjsfrozen Markdown-closure counts updated (edges 307 → 308) for the one new cross-reference link.git-hooks.md,verification-environment.md, or runtime/executable behavior; no companion case study.Change Type
Test and Review Evidence
npm cinpx vitest run test/skills-docs/doc-link-graph.test.mjsnpx vitest run test/skills-docs/npm testManual or visual evidence: full diff reviewed; new sections follow the existing reference style (bold run-in leads, tables, anti-pattern list) and reuse the four-valued verdict domain.
Risk and Recovery
AI Involvement
Checklist
AGENTS.md,CONTRIBUTING.md, and the relevant canonical-owner guidance.CHANGELOG.md.Note on the two unchecked items: no shipped files or dependencies changed (docs + one test snapshot constant), and there is no user-facing behavior or compatibility change, so
CHANGELOG.mdwas intentionally left untouched per its "observable behavior and compatibility" scope.