AI Workflows
Specifications
| Convention | Spec |
|---|---|
| SKILL.md | Agent Skills Specification |
| AGENTS.md | agents.md Standard |
| llms.txt | llms.txt Specification |
Skills vs Docs
- Skills (
skills/<name>/SKILL.md) = executable agent instructions — how to do things (conventions, workflows, tool usage) - Docs (
docs/) = project documentation — what was decided and why (rfcs/, guides/, plans/, runbooks/, architecture/) - Both exist in every project — neither replaces the other
- Every repo should also have: README.md (humans) + AGENTS.md (AI) + llms.txt (LLM discovery)
- Root-level standard files (README, AGENTS.md, CONTRIBUTING, LICENSE, CODEOWNERS, llms.txt) stay at root — everything else goes in
docs/
AI Tools
| Tool | Usage |
|---|---|
| Claude Code (opus) | Primary AI coding assistant — architecture, refactoring, multi-file edits |
| OpenAI Codex | Alternative AI assistant for code generation |
| Google Gemini CLI | Google’s AI for research and code tasks |
| GitHub Copilot | Inline code completion in Neovim |
Claude Code Configuration
Claude Code is configured per-project via CLAUDE.md files at the repository root. These files define:
- Project context — tech stack, directory structure, key patterns
- Coding conventions — naming, formatting, commit style
- Behavioral rules — what to avoid, how to handle edge cases
- Memory — persistent notes stored in
.claude/directories
Best Practices
- Keep
CLAUDE.mdconcise — it’s loaded into every conversation - Use separate memory files for detailed topic notes
- Reference project-specific tools and commands
- Define verification steps (build, test, lint)
AGENTS.md Standard
Every repo MUST have AGENTS.md at root.
Template Structure
- Identity — project name, one-line description
- Architecture — high-level component overview
- Commands — build, test, lint, run
- Code Style — language-specific conventions
- Extension Guide — how to add features
Do NOT include static directory trees or file tables — structure goes stale. Instead, instruct agents to use tree and ripgrep/ag to discover project layout on the fly.
Skills Standard
Every repo with a CLI or library should have skills/<name>/SKILL.md:
- Frontmatter:
name,description,argument-hint(for CLI tools) - Content: step-by-step guidance for using or extending the project
- Skills live in
skills/<name>/SKILL.mdwithin each repo
AI-Powered Git Workflows (sr)
sr provides AI-powered git commands with multi-backend support (--backend {claude|copilot|gemini}):
| Command | Purpose |
|---|---|
sr commit | Generate atomic conventional commits from staged/unstaged changes |
sr rebase | AI-powered interactive rebase (reword, squash, reorder) |
sr review | Code review of staged/branch changes with severity feedback |
sr pr | Generate PR title + body from branch commits |
sr branch | Suggest conventional branch name |
sr explain | Explain recent commits |
sr ask | Freeform Q&A about the repo |
# Stage changes, then generate commit
git add -p
sr commit
# Or commit all changes with context
sr commit -M "refactored auth flow"
# AI-powered rebase of last 5 commits
sr rebase --last 5
Global flags: --backend, --model, --budget (Claude only, default $0.50), --debug.