Skip to content

Improve AI agent readiness with comprehensive documentation#769

Open
gmcculloug wants to merge 1 commit intoRedHatInsights:masterfrom
gmcculloug:improve-agent-readiness
Open

Improve AI agent readiness with comprehensive documentation#769
gmcculloug wants to merge 1 commit intoRedHatInsights:masterfrom
gmcculloug:improve-agent-readiness

Conversation

@gmcculloug
Copy link
Copy Markdown
Contributor

This PR establishes a layered documentation system for AI-assisted development.

What Changed

New Files

Domain-specific guideline files (docs/*-guidelines.md):

  • security-guidelines.md — Permission format, role flags, privilege escalation prevention, CI security checks
  • api-contracts-guidelines.md — JSON schema contracts, field specifications, naming conventions
  • database-guidelines.md — Role seeding, version bumping, KSL schema patterns, tenant isolation
  • testing-guidelines.md — CI validation pipeline, local testing commands, pre-PR checklist
  • integration-guidelines.md — Deployment pipeline, GitHub actions, ConfigMap generation

Agent documentation:

  • AGENTS.md — Repository layout, cross-cutting conventions, common pitfalls, PR workflow (agent-agnostic)
  • CLAUDE.md — Claude Code-specific commands and behavioral preferences (imports AGENTS.md)
  • CONTRIBUTING.md — Contribution workflow guide for humans and agents

Configuration:

  • .coderabbit.yaml — Points CodeRabbit to guideline files for automated code review enforcement

Modified Files

  • README.md — Enhanced structure, added documentation index, improved organization

Why This Matters

This layered documentation system enables:

  1. Consistent agent behavior — Claude, Cursor, CodeRabbit, and other AI tools follow the same conventions
  2. Automated PR reviews — CodeRabbit enforces guideline files during code review
  3. Reduced onboarding friction — New contributors (human and AI) have clear, comprehensive guidance
  4. Better knowledge preservation — Repo-specific conventions are documented, not just tribal knowledge

Documentation Structure

CLAUDE.md (Claude-specific) → @AGENTS.md (all agents) → docs/*-guidelines.md (deep domain expertise)
  • README.md — Front door for humans and agents (high-level overview)
  • CONTRIBUTING.md — How to contribute
  • AGENTS.md — Cross-cutting conventions, common pitfalls, docs index
  • CLAUDE.md — Claude Code build commands and preferences
  • docs/*-guidelines.md — Deep domain-specific rules (security, testing, database, etc.)

https://redhat.atlassian.net/browse/RHCLOUD-46839

🤖 Generated with Claude Code

This PR establishes a layered documentation system for AI-assisted development:

**Created:**
- Domain-specific guideline files (docs/*-guidelines.md) for security, API contracts, database, testing, and integration
- AGENTS.md: Agent onboarding doc with cross-cutting conventions and docs index
- CLAUDE.md: Claude Code-specific layer with build commands and behavioral preferences
- CONTRIBUTING.md: Contribution workflow guide for humans and agents
- .coderabbit.yaml: CodeRabbit configuration pointing to guideline files

**Enhanced:**
- README.md: Improved structure, added documentation index, better organized

https://redhat.atlassian.net/browse/RHCLOUD-46839
@gmcculloug gmcculloug requested a review from lpichler April 22, 2026 20:37
@gmcculloug
Copy link
Copy Markdown
Contributor Author

Bump @lpichler Is it possible to review before pto? Or assign to some who should review.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant