Skip to content

AGENTS.md

Latest commit

 

History

History
103 lines (67 loc) · 2.73 KB

AGENTS.md

File metadata and controls

103 lines (67 loc) · 2.73 KB
Raw

AGENTS.md

Guidance for AI agents and human contributors working on this 11ty-based personal website.

Project Structure

  • src/ - Content (markdown, templates, posts)
  • _config/ - Modular Eleventy configuration
  • _site/ - Build output (not committed)
  • .eleventy.js - Main config

Development

npm start              # Local dev server with live reload
npm run build:prod     # Production build

Environment: Node 20 (.nvmrc), optional ELEVENTY_ENV and WEBMENTIONS_TOKEN

Git Workflow

Conventional Commits

Follow Conventional Commits format: <type>: <description>

Types: feat, fix, docs, style, refactor, test, chore

Examples: feat: add dark mode, docs: update Readme.md

Atomic Commits

  • One logical change per commit
  • Independently revertible
  • Avoid mixing unrelated changes

Coding Standards

Do:

  • Use modular config (_config/ pattern)
  • Leverage existing 11ty plugins
  • Maintain semantic HTML
  • Follow established patterns

Don't:

  • Commit build artifacts (_site/)
  • Hardcode environment values
  • Break existing content
  • Add unjustified dependencies

Collaboration

Quality Over Quantity:

  • Prefer thorough, well-tested changes over rapid iterations
  • Self-review before committing
  • Align with project goals

Communication:

  • Explain why, not just what in commit messages
  • Document non-obvious decisions
  • Update docs when behavior changes

Agentic Development

For AI Agents:

  • Understand context first - review existing files
  • Make minimal, necessary changes only
  • Test locally with npm start
  • Respect established conventions
  • Validate builds succeed

Pre-Commit Checklist:

  1. Site builds without errors
  2. Content renders correctly
  3. No broken links/assets
  4. Conventional commit format followed

Compound Engineering

Session Evaluation:

At the end of each session, agents should evaluate their work and propose improvements:

  1. AGENTS.md Improvements: Review if new patterns or conventions emerged that should be documented
  2. New Agent Needs: Identify if specialized agents in .github/agents/ would improve efficiency (e.g., testing-agent, docs-agent, security-agent)
  3. Skill Gaps: Suggest specific skills for .github/skills/ that would benefit future work (e.g., 11ty-optimization, markdown-linting, accessibility-checks)

Evaluation Questions:

  • What repetitive tasks could be automated by a specialized agent?
  • What knowledge would benefit future agents working on this codebase?
  • What conventions or patterns should be added to AGENTS.md?
  • What tools or workflows could be standardized in skills files?

Keep this document updated as the project evolves.