Trim AGENTS.md to align with best practices (369 → 122 lines)#39
Trim AGENTS.md to align with best practices (369 → 122 lines)#39
Conversation
The wg-agentic-sdlc repo-scaffolding best practices recommend AGENTS.md files stay under 150 lines, citing ETH Zurich research showing agent adherence drops with context file length. The prior AGENTS.md was 369 lines, with most content duplicating what each workflow's guidelines.md and skills/*.md already provide in the AI execution chain. Removed sections verified as duplicated: - Workflow-Specific Notes (covered in each workflow's guidelines.md) - Prerequisites per workflow (in each workflow's guidelines.md/skills) - Artifact Management per-workflow listing (in each workflow's docs) - Common Commands (generic tooling agents already know) - Verbose Shared Resources explanation (files are self-documenting) Relocated unique content: - Claude Code installation internals moved to CONTRIBUTING.md Added: - One-sentence recipe concept explanation after the _shared/ tree - Alphabetized workflow list - README.md annotation noting it contains prerequisites and artifacts Validated by running a fresh-session test confirming an agent with only the trimmed AGENTS.md could navigate the repo, understand the recipe system, identify relevant workflows, and follow path conventions. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
Warning Rate limit exceeded
You’ve run out of usage credits. Purchase more in the billing tab. ⌛ How to resolve this issue?After the wait time has elapsed, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout. Please see our FAQ for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (3)
WalkthroughThis PR updates workflow documentation standards and installation guidance. ChangesWorkflow Documentation and Installation Guidance
🎯 2 (Simple) | ⏱️ ~8 minutes 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@AGENTS.md`:
- Around line 77-84: The CI’s AGENTS structure validator is failing because
AGENTS.md no longer includes the workflow-specific notes/artifact sections the
validator expects; update the validator (or reintroduce a minimal compatible
section) so it accepts the trimmed AGENTS contract — locate the validator script
(the AGENTS/lint step in your CI workflow) that enforces headings like "##
Installation" and "## Development" and change its rules to allow the new format
(or add a minimal "Workflows/Artifacts" compatibility block to AGENTS.md at the
same semantic location as the removed sections) so entries around the "##
Installation" and "## Development" headings pass CI.
In `@CONTRIBUTING.md`:
- Around line 90-92: Replace the home-absolute path "~/.claude/skills/" in the
second list item with a path-agnostic phrase (e.g., "global Claude skills
directory") and ensure the alternate project-level reference remains a relative
path (".claude/skills/"); update the sentence so it reads something like
"Symlinks workflows into the global Claude skills directory (or .claude/skills/
for project-level) for slash command discovery" and remove any other
home-absolute references to comply with the relative-path rule.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 53fd0541-8b01-467f-9a75-7bd424a0bb74
📒 Files selected for processing (2)
AGENTS.mdCONTRIBUTING.md
Update validate-structure.sh to match the trimmed AGENTS.md format: - Remove checks for per-workflow artifact entries and workflow-specific notes sections in AGENTS.md (content now lives in each workflow's docs) - Add new check ensuring each workflow's README.md documents its artifact paths, catching the same "forgot to document artifacts" problem at the source rather than in a summary Add artifacts section to ai-ready/README.md to satisfy the new check (no .artifacts/ output — writes directly to the target project). Replace home-absolute path ~/.claude/skills/ in CONTRIBUTING.md with path-agnostic wording per CodeRabbit feedback. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Summary
Motivation
Cross-pollination analysis between wg-agentic-sdlc and this repo identified that our AGENTS.md exceeded the recommended 150-line limit (hard cap 300). ETH Zurich research shows agent adherence drops with context file length. Duplication analysis confirmed that Workflow-Specific Notes, Prerequisites, Artifact Management, and Common Commands were already documented in each workflow's own guidelines.md and skills/*.md files — the files AI agents actually read during execution.
What changed
Deleted (duplicated in workflow docs):
Relocated:
Added:
_shared/tree(prerequisites, artifacts, usage)Validation
Test plan
bugfix /assessorcode-review /start) against a target project to confirm the agent can still orient itselfskill-reviewer /reviewagainst a workflow in this repo to confirm it can discover the architecture🤖 Generated with Claude Code