1ai-skills: Instructions file for Claude Code

AGENTS.md — 1ai-skills

MANDATORY PROCESS (8 Steps — No Skipping)

Every task follows this sequence. No exceptions.

1. AUDIT — Read existing code. Understand current state.
2. THINK — Understand WHY. Intent vs literal.
3. BRAINSTORM — ≥3 approaches. Score options.
4. PLAN — Decompose. Risks. Rollback plan.
5. EXECUTE — Build. TDD when possible.
6. TEST — Run all tests. Break it first.
7. VERIFY — Prove with literal output.
8. REVIEW — Read your own diff before committing.

Full details: ~/.1ai/core/PROCESS.md (auto-injected by hooks)

This repo

Production-ready AI agent skill library — 1372 skills across 19 categories with anti-rationalization tables, code examples, and verification checklists. Stack: Node.js / Python / YAML skills Domain: Agent skill library — NOT a product repo. Do not add features here unless adding new skills.

Rules — thin loader, no submodule

Engineering rules are enforced by machine-level loaders when setup-dev.sh has been run:

• Claude Code: SessionStart hook injects ~/.1ai/core/RULES.md
• OpenCode: plugin injects ~/.1ai/core/RULES.md
• OMP: wrapper appends ~/.1ai/core/RULES.md to launch sessions

Primary rules file:

cat ~/.1ai/core/RULES.md

Pre-ship gate:

cat ~/.1ai/core/GATE.md

What this repo is

1ai-skills provides domain-specific playbooks for AI agents. Each skill has:

• YAML frontmatter (name, description, domain, tags)
• Anti-Rationalization Table (prevents cutting corners)
• Workflow (step-by-step)
• Code examples (Python/JS/Bash)
• Verification checklist

1ai-rules = HOW to code. 1ai-skills = WHAT to do.

Finding and using skills

# Find by category
ls ~/projects/1ai-skills/development/    # TDD, debugging, code review, PRD
ls ~/projects/1ai-skills/devops/         # Docker, K8s, CI/CD
ls ~/projects/1ai-skills/integrations/   # Stripe, Firebase, Supabase, GitHub
ls ~/projects/1ai-skills/cybersecurity/  # 786 security skills
ls ~/projects/1ai-skills/trading/        # Crypto, DeFi, Polymarket

# Search via brain MCP (if 1ai-hub is running)
# gbrain_search or vilona_brain_search: "stripe integration skill"

# Search SKILLS.json directly
python3 -c "
import json
s = json.load(open('SKILLS.json'))
hits = [sk for sk in s.get('skills', []) if 'stripe' in sk.get('name','').lower()]
for h in hits: print(h['name'], '-', h.get('description','')[:80])
"

# Read a skill
cat ~/projects/1ai-skills/integrations/stripe-integration/SKILL.md