[instructions] Sync instruction files with release v0.85.0

[github-actions]

Summary

Ran the instructions-janitor Release/Change, Size, Duplication, and Accuracy audits (baseline: last sync commit 97b00a63, covering 90 commits through 80b45ff3, release v0.85.0).

Files changed

• .github/aw/designer.md — 416 → 394 lines
• .github/aw/github-agentic-workflows.md — 160 → 161 lines
• .github/aw/update-agentic-workflow.md — 121 → 122 lines
• .github/aw/configure-agentic-engine.md — 177 → 177 lines
• .github/aw/syntax-engine.md — 87 → 87 lines
• .github/aw/safe-outputs-runtime.md — 298 → 300 lines

What changed and why

• Size fix: designer.md regrew past the 400-line target (416 lines) after a new "Phase 1b: Repository Survey" section was added that fully re-stated the survey signal list already owned by the new maintainer.md (added in Add repository maintainer sample prompt #50413/Favor shared workflow engine configuration #50422 for recurring repository-maintenance workflows). Replaced the restated section with a short pointer to maintainer.md, and collapsed designer.md's redundant dual "In-repo references" / "Portable HTTPS references" lists (which mirrored each other 1:1) into a single list plus a one-line prefix-substitution note — no other .github/aw/*.md file uses that dual-list pattern. Net: 416 → 394 lines, duplication removed.
• Discoverability: maintainer.md (208 lines, new) was only linked from designer.md and create-agentic-workflow.md. Added it to github-agentic-workflows.md's compact reference table and update-agentic-workflow.md's topic-file list so it's reachable the same way campaign.md/experiments.md are.
• Accuracy — engines: configure-agentic-engine.md's shared-engine example list named OpenCode/Goose/Aider/Cursor/Kiro but not Crush (.github/workflows/shared/crush.md, added in Implementing shared agentic workflow with charmbracelet/crush #50430) — added it. syntax-engine.md only called out opencode as an importable shared engine; generalized that line to cover all shared CLI engines and point at configure-agentic-engine.md instead of hardcoding one example (reduces future staleness as engines are added/removed — antigravity's removal in Removing support for the antigravity agentic engine #50460 was already reflected correctly).
• Accuracy — safe outputs: cross-checked SafeOutputsConfig (pkg/workflow/safe_outputs_config_types.go) against the safe-outputs docs; found report-failed-jobs (added in feat: report failed non-builtin jobs as issues from conclusion job #49959, controls whether failed custom jobs: are reported as issues from the conclusion job) had no documentation anywhere. Added it next to the related report-failure-as-issue/failure-issue-repo fields in safe-outputs-runtime.md.

Documentation commits reviewed

git log/compare-API pass over 97b00a63...80b45ff3 (90 commits), with focused diff review of: #50413/#50422 (maintainer.md + designer.md/create-agentic-workflow.md survey integration), #50278/#50244/#50234 (agentfinder removal from skills.md), #50460 (antigravity engine removal), #50430 (Crush shared engine), #49959 (report-failed-jobs), plus the configure-agentic-engine.md rewrite toward shared-markdown engine definitions.

Audits performed

• Release/change audit: compared last sync commit 97b00a63 (2026-08-04) to HEAD 80b45ff3 via the GitHub compare API (90 commits). Identified 8 touched .github/aw/*.md files plus actions-lock.json (routine digest bump, no action needed).
• Size audit: wc -l .github/aw/*.md — only designer.md exceeded 400 lines (416); fixed to 394. All other files, including the newly added maintainer.md (208), are within target.
• Duplication audit: found and removed the designer.md/maintainer.md survey-content duplication and designer.md's redundant dual reference-list pattern (unique to that file — verified via grep across all .github/aw/*.md).
• Accuracy audit: spawned an independent Explore agent to cross-check syntax-engine.md, safe-outputs*.md, syntax*.md/create-agentic-workflow.md, and configure-agentic-engine.md against pkg/workflow/agentic_engine.go, pkg/workflow/safe_outputs_config_types.go, pkg/parser/schemas/main_workflow_schema.json, and pkg/workflow/data/engines/. Confirmed the recent antigravity-removal and shared-engine-definition rewrites are accurate; found and fixed the one real gap (report-failed-jobs) plus the Crush/opencode-only engine-list gaps above.

Confirmation

• Duplicated content identified and removed (designer.md ↔ maintainer.md survey section; designer.md's dual reference lists)
• Every edited .github/aw/*.md file is under its target limit (largest: designer.md at 394/400 lines)

Generated by 🧹 Instructions Janitor · sonnet50 · 321.7 AIC · ⌖ 6.76 AIC · ⊞ 5.2K · ◷

• expires on Aug 7, 2026, 1:35 AM UTC-08:00

---

Run: https://github.com/github/gh-aw/actions/runs/30998678554

Generated by 👨‍🍳 PR Sous Chef · gpt54 · 8.91 AIC · ⌖ 7.13 AIC · ⊞ 8.3K · ◷
_{Comment /souschef to run again}

[github-actions]

🎯 Thanks for keeping the instruction documentation in sync with v0.85.0!

Your audit-driven approach looks solid:

• Duplication removal: Collapsed the redundant survey-section duplication between designer.md and maintainer.md, plus the dual reference-list pattern in designer.md.
• Accuracy fixes: Added the missing Crush engine reference, generalized the engine-list approach to reduce future staleness, and documented the previously undocumented report-failed-jobs config field.
• Discoverability: Added maintainer.md links to the reference tables in github-agentic-workflows.md and update-agentic-workflow.md so it's discoverable the same way as campaign.md/experiments.md.

The PR is well-scoped to documentation, the changes are substantive and well-reasoned, and the audit confirms accuracy across the engine and safe-outputs config spaces. This looks ready for review! 🚀

Generated by ✅ Contribution Check · auto · 49.3 AIC · ⌖ 5.05 AIC · ⊞ 8.8K · ◷

[gh-aw-bot]

@copilot Please take another pass on this PR.

• The latest activity looks positive and there are no failed checks in the compact queue snapshot.
• Please refresh the branch if needed, confirm final readiness, and run the pr-finisher skill before handing back.

Generated by PR Sous Chef · run https://github.com/github/gh-aw/actions/runs/30994598457

Generated by 👨‍🍳 PR Sous Chef · gpt54 · 6.23 AIC · ⌖ 12.1 AIC · ⊞ 8.3K · ◷
_{Comment /souschef to run again}

[gh-aw-bot]

@copilot Please take another pass on this PR.

• Refresh the branch if needed and confirm final readiness.
• Run the pr-finisher skill before handing back.

Generated by PR Sous Chef · run https://github.com/github/gh-aw/actions/runs/30998678554

Generated by 👨‍🍳 PR Sous Chef · gpt54 · 8.91 AIC · ⌖ 7.13 AIC · ⊞ 8.3K · ◷
_{Comment /souschef to run again}