docs(#5560): align suite-timeout authorization scope

[fullsend-ai-coder]

Summary

Align the authorization-source language in the suite-timeout policy so both paragraphs reference the same scope. Paragraph 2 previously said "linked issue or PR description" while paragraph 1 said "issue/PR comment" — these are distinct scopes in GitHub's data model. Updated paragraph 2 to say "linked issue or a PR comment" to match.

Related Issue

Fixes #5560

Changes

• Updated docs/contributing/go-code.md: changed "Explicit human authorization in the linked issue or PR description" to "Explicit human authorization in the linked issue or a PR comment" in the reviewer-facing paragraph

Testing

• Verified both paragraphs now use consistent authorization-source language via grep
• No code changes — documentation only

---

Closes #5560

Post-script verification

• Branch is not main/master (agent/5560-fix-timeout-auth-scope)
• Secret scan passed (gitleaks — 331d7498ce2dc4eae90a24883749e3d349fa1da9..HEAD)
• PR body secret scan passed (gitleaks — no-git)
• Pre-commit hooks passed (authoritative run on runner)
• Tests ran inside sandbox

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 7:13 AM UTC · Completed 7:27 AM UTC
Commit: 9a18708 · View workflow run →

[github-actions]

Site preview

Preview: https://910c9aae-site.fullsend-ai.workers.dev

Commit: 9a187087d77842bc0e5936a5317e7582da26b7ba

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 7:23 AM UTC · Completed 7:35 AM UTC
Commit: 9a18708 · View workflow run →

[fullsend-ai-review]

Review skipped — this PR is already merged.

The /fs-review command only reviews open pull requests.

_{Posted by fullsend post-review check}

[fullsend-ai-retro]

Retro: PR #5561 — docs(#5560): align suite-timeout authorization scope

Workflow Overview

This PR is the product of a fully automated retro→triage→code pipeline, with human involvement only at the approval/merge step. The chain started when the retro agent analyzed PR fullsend-ai/fullsend#5527 (which added a suite-timeout policy) and identified an unresolved qodo-code-review finding about inconsistent authorization-scope language between two paragraphs of the new policy. The retro agent filed issue fullsend-ai/fullsend#5560, the triage agent confirmed it as low-priority documentation ready-to-code, and the code agent produced a correct 1-line fix in PR #5561.

Timeline

Time (UTC)	Event
07:01	Retro agent files issue #5560 from PR #5527 analysis
07:02–07:06	Triage agent processes issue (run 30074080387)
07:06–07:12	Code agent implements fix, opens PR #5561 (run 30074360202)
07:12	PR opened; first review dispatch triggered
07:16	Human (rh-hemartin) approves, adds to merge queue; second review dispatch triggered
07:20	PR merged; third review dispatch triggered
07:27	First review agent run completes ($2.78, 15 min) — detects PR already merged, skips posting

What Went Well

• Retro→code pipeline executed flawlessly. The retro agent correctly identified a qodo finding that was ignored during PR docs(agents): forbid unsolicited e2e suite timeout bumps #5527's review, filed a well-structured issue with a specific recommended fix, and the code agent produced the exact prescribed change. Total automated pipeline time from issue to PR: 11 minutes.
• Code agent quality was high. The 1-line change correctly aligned paragraph 2's authorization scope language ("PR description" → "a PR comment") to match paragraph 1's broader scope.
• Fast human turnaround. The reviewer approved and merged within 8 minutes of the PR opening.

Known Inefficiencies (All Tracked by Existing Issues)

Review agent race condition. The review agent took ~15 minutes for a 1-line docs change. The human approved and merged 7 minutes before the review completed. The agent correctly detected the merged state and skipped posting, but $2.78 of compute was consumed with no output. Three separate review runs were dispatched (PR opened, approval event, merge event). This pattern is well-tracked:

• Review completing after merge: #2388, #4509, #4768, agents#337
• Preventing unnecessary dispatches: #3010, #4070
• New evidence: review run 30074680504 spent $2.78 (4 sub-agents + Opus challenger, 977K cached read tokens, 33K output tokens) on a 1-line docs wording change that was merged 7 minutes before the review completed.

Disproportionate review effort for trivial changes. The review agent allocated its full sub-agent fleet (correctness, style-conventions, intent-coherence, docs-currency, plus Opus challenger) for a single-line documentation wording fix. Well-tracked:

• agents#373: skip irrelevant review dimensions for docs-only PRs
• agents#301: lightweight path for trivial/mechanical changes
• #5157: scale review detail to change complexity
• #1900: fast path for trivial docs/prompt-only changes
• #3240: fast-path triage for trivially safe bot PRs

Original review gap on PR #5527. The review agent on PR #5527 did not independently detect the authorization-scope inconsistency that qodo caught. This gap in internal-consistency detection for rule/guidance documents is tracked by #3901 and agents#265.

Assessment

No novel proposals. This was a well-functioning workflow demonstrating the retro agent's value in catching review findings that fell through the cracks. All identified inefficiencies are covered by existing open issues with good specificity. The main takeaway is a positive one: the retro→triage→code pipeline produced a correct fix autonomously in 11 minutes, and the only human involvement was a quick approval/merge.