🧠 Personal AI Skill Library

A vendor-neutral, multi-agent skill library for AI-powered software engineering

---

🌐 Tiếng Việt

---

🎯 Keep repeatable AI workflows in one place. Separate skills from knowledge. Work across any agent.

🏗️ Design Goals

	Goal
🔹	Skills are small, specific, and reusable — with modes instead of separate files
🔹	Knowledge is stored separately from skills
🔹	Adapters are generated from one source — not hand-maintained
🔹	Works without any custom plugin runtime (general skills)
🔹	Anionzo ecosystem skills extend the core with structured workflows

---

🎯 Skill Catalog

🔧 Tool Dependencies

Tool	Type	Purpose
br / bv	Core	Beads planning and visualization
Agent Mail MCP	Core	Inter-agent messaging for swarming
gkg	Optional	Codebase intelligence
CASS / CM	Optional	Session search and context

See tools/ for full tool documentation.

16 skills in three categories: Core Chain, Anionzo Chain, and Standalone.

═══════════════════════════════════════════════════

🧩 Core Chain

Primary workflow — works in any project

═══════════════════════════════════════════════════

using-skills → brainstorming → research → planning → feature-delivery
     (router)        (if vague)    (before impl)     │
                                                  ┌──┴──┐
                                                  ▼     ▼
                                            debug  docs-writer
                                                  │
                                             code-review
                                                  │
                                                commit
                                                  │
                                               extract

	Skill	Purpose	Modes
🧭	using-skills	Route a request to the right skill and mode	router
💡	brainstorming	Explore ideas, lock decisions, write spec	quick · spec · deep-explore
🔎	research	Explore codebase, onboard repo, deep-scout feature discovery, upgrade prompts, codebase intel	quick-search · repo-bootstrap · deep-scout · prompt-upgrade · codebase-intel
📐	planning	Research → plan → validation gate before any code is written	full pipeline + Phase 8 validation gate
🚀	feature-delivery	Implement, test-first, or refactor	standard · tdd · refactor
🐛	debug	4-phase systematic debugging with root cause investigation	+ anionzo ecosystem extensions
📖	docs-writer	Create or update documentation from verified source behavior	prompt-only · docs-execution · prompt+execution
🔍	code-review	Give reviews, receive reviews, verify before claiming done	verification-gate · giving · receiving
📝	commit	Create clean conventional commits with staged review	—
🧬	extract	Extract learnings, session handoff, deep compounding, dream consolidation	handoff · extract · compound · dream

---

═══════════════════════════════════════════════════

⚙️ Anionzo Chain

Structured multi-agent workflow — anionzo ecosystem only

═══════════════════════════════════════════════════

using-anionzo → brainstorming → research → planning → swarming → reviewing → extract
   (bootstrap)      (deep-explore)   (+ validation gate)   (workers)    (5-agent verify)    (compound)

	Skill	Purpose	Position
⚙️	using-anionzo	Bootstrap anionzo projects: onboarding, STATE.md, go-mode pipeline	entry point
🐝	swarming	Orchestrate parallel worker agents with rescue coordination	phase 5 of 9
📋	reviewing	Post-execution verification: 5 specialist agents + artifact checks + human UAT	phase 7 of 9

---

═══════════════════════════════════════════════════

🔧 Standalone Skills

Secondary — load only when the task matches

═══════════════════════════════════════════════════

Standalone skills remain available, but they are intentionally secondary to the Core and Anionzo chains in this repo's top-level narrative.

	Skill	Purpose
🎨	animated-landing-pages	Motion-first landing page with AI-generated visuals
📚	book-sft-pipeline	Fine-tune models on book style: ePub → SFT dataset → LoRA training
🛠️	writing-anionzo-skills	Create or edit anionzo skills using TDD methodology

---

Skill Modes at a Glance

┌──────────────────┬──────────────────────────────────────────────────────┐
│     SKILL         │  MODES                                                   │
├──────────────────┼──────────────────────────────────────────────────────┤
│ brainstorming    │ quick · spec · deep-explore                              │
│ research         │ quick-search · repo-bootstrap · deep-scout · prompt-upgrade │
│                  │   · codebase-intel                                        │
│ feature-delivery │ standard · tdd · refactor                               │
│ docs-writer      │ prompt-only · docs-execution · prompt+execution          │
│ code-review      │ verification-gate · giving · receiving                 │
│ extract          │ handoff · extract · compound · dream                   │
└──────────────────┴──────────────────────────────────────────────────────┘

Skill	Mode	When to use
brainstorming	quick	Lock direction only — no output artifact
brainstorming	spec	Write full spec: FR/NFR/ACs/Given-When-Then
brainstorming	deep-explore	Socratic dialogue + locked decisions + CONTEXT.md (anionzo)
research	quick-search	Targeted lookup in known repo
research	repo-bootstrap	Onboard to unfamiliar repo
research	deep-scout	High-risk feature: map stack + local reuse + upstream + official docs (HARD-GATE: no code before brief)
research	prompt-upgrade	Refine rough prompt into execution-ready instruction
research	codebase-intel	Use gkg MCP tools for architecture snapshot
feature-delivery	standard	Normal feature implementation
feature-delivery	tdd	Test-first: red-green-refactor before production code
feature-delivery	refactor	Restructure code without changing behavior
docs-writer	prompt-only	Return upgraded prompt, no doc execution
docs-writer	docs-execution	Directly update docs from live repo
docs-writer	prompt+execution	Both: return prompt and execute
code-review	verification-gate	Iron law: no claim without fresh evidence
code-review	giving	Review diffs, PRs, commit ranges
code-review	receiving	Respond to review feedback
extract	handoff	Session near context limit — compress state for next session
extract	extract	Capture durable learnings from completed task
extract	compound	Post-merge deep analysis: 3 parallel subagents (anionzo)
extract	dream	Consolidation pass over accumulated learnings (anionzo)

---

📖 Research Highlights

Source	Key Pattern
🏛️ anthropics/skills	Minimal, portable skill packaging
⚡ obra/superpowers	Brainstorm → plan → execute → verify workflow
🧩 affaan-m/everything-claude-code	Layered model: skills, rules, memory, adapters
🗃️ knowns-dev/knowns	Separate skills from knowledge; generate platform files
📦 hoangnb24/skills	Workflow-first skill design with router and output contracts
🧩 anionzo/skills	Anionzo ecosystem: bead graph, swarm orchestration, compounding

---

🚀 Quick Start

npx @anionzo/skill

For a reusable local CLI:

cd /path/to/skill
npm link

Then inside any project:

anionzo init
# or
anionzo skill init

Non-interactive:

anionzo init --yes
anionzo init --platform opencode,claude,copilot,gitignore
npx @anionzo/skill --platform opencode,agents

• Windows: Uses PowerShell automatically (.cmd + .mjs + .ps1)
• Linux/macOS: Uses bash script

Version 2.0.0 required.

npm cache clean --force
npx --yes @anionzo/skill

After installation, open your agent and say:

• Use the using-skills router for this task
• Use the anionzo go-mode pipeline for this feature
• Help me understand this repo first
• Plan this feature, then implement it

---

📦 Install via npm

# Recommended
npx @anionzo/skill

# Quiet first-run
npx -y --loglevel=error @anionzo/skill

# Add as project dependency
npm install @anionzo/skill

🔌 Agent Integration

This repo is the source of truth. Generated files are delivery artifacts only.

Agent	Copy from	Copy to
🤖 Claude Code	generated/CLAUDE.md	CLAUDE.md
⚡ OpenCode	generated/OPENCODE.md	OPENCODE.md
💎 Gemini CLI	generated/GEMINI.md	GEMINI.md
🔧 Generic	generated/AGENTS.md	AGENTS.md
🐙 GitHub Copilot	generated/copilot-instructions.md	.github/copilot-instructions.md

---

➕ Create A New Skill

# 1. Scaffold from template
cp -r templates/ skills/<new-skill>/

# 2. Edit the files
vim skills/<new-skill>/SKILL.md
vim skills/<new-skill>/meta.yaml

# 3. Validate
bash scripts/validate-skills

# 4. Regenerate platform files
bash scripts/sync-platform-files

For anionzo ecosystem skills, also read skills/writing-anionzo-skills/SKILL.md for TDD-for-skills methodology.

---

⚙️ Commands

Command	Purpose
bash scripts/validate-skills	Check all skills have required files and keys
bash scripts/sync-platform-files	Generate platform instruction files

---

📋 Recommended Customization Order

1. 🥇 knowledge/global/engineering-principles.md
2. 🥈 knowledge/global/review-heuristics.md
3. 🥉 knowledge/global/debugging-patterns.md
4. 🎯 skills/using-skills/SKILL.md — the router, customize routing for your stack
5. 🔌 Adapter output for your two most-used agents

---

🤝 Contributing

We welcome contributions! See CONTRIBUTING.md for:

• ➕ How to add a new skill
• ✏️ How to edit existing skills
• 📚 How to contribute knowledge
• 🔄 Pull request process and conventions

---

📌 Notes

• generated/ is gitignored — regenerate after any skill/knowledge change
• General skills (no deps) work in any project without plugins
• Anionzo ecosystem skills require .anionzo/ onboarding (run anionzo init first)
• Standalone skills are secondary to the Core and Anionzo chains
• All skills follow the Shared Output Contract: Goal/Result → Key Details → Next Action

---

Built with ❤️ for AI-assisted software engineering