use-skill

name	use-skill
description	Meta-skill that catalogs all available skills, explains when and how to use each one, and auto-discovers new skills. ALWAYS read this skill first before starting work to understand what tools are available. Use this as the single source of truth for skill selection.

Use-Skill: Skill Catalog & Usage Guide

ALWAYS READ THIS SKILL FIRST before starting any work. This is your single source of truth for what skills are available and when to use them.

Quick Start

# Scan for all available skills
bash skills/use-skill/scripts/scan-skills.sh

Currently Installed Skills

🧠 Core Agent Capabilities

Skill	When to Use	Key Features
agent-browser	Headless browser automation outside OpenClaw's built-in browser	Navigate, click, fill forms, screenshots, PDFs, network interception, console inspection
codebase-cartographer	Understanding large codebases	Rust-based agentlens CLI, generates .agentlens/ docs, finds TODOs, symbols, dependencies
agent-swarm-workflow	Multi-agent parallel implementation	NTM + Agent Mail + BV coordination, exact prompts for swarm execution
council	Complex decisions needing multiple expert perspectives	Multi-persona deliberation (architect, analyst, security), 3-turn debate structure
autonomous-skill-orchestrator	Tasks requiring autonomous execution without babysitting	Freeze intent → plan → execute → loop, strict guardrails

🔧 Development & Code Quality

Skill	When to Use	Key Features
skill-creator	Creating new skills or updating existing ones	init_skill.py, package_skill.py, validation
memory-hygiene	Vector memory cleanup and optimization	Audit LanceDB, wipe/reseed, disable auto-capture
ai-cron-gen	Generating cron expressions from natural language	"Every Monday at 3pm" → "0 15 * * 1"

📋 Project Management

Skill	When to Use	Key Features
beads-workflow	Task tracking with bd (beads)	Create, update, sync tasks with git
agent-mail	Multi-agent coordination via mail system	Inbox/outbox, file reservations, threaded messaging

🌐 External Integrations

Skill	When to Use	Key Features
moltbook-interact	Reading from Moltbook agent social network	Browse posts, analyze (requires API credentials)

Skill Selection Decision Tree

Starting a task?
├── Need browser automation?
│   └── Use: agent-browser
├── Working with large codebase?
│   └── Use: codebase-cartographer
├── Complex decision with trade-offs?
│   └── Use: council
├── Want to run autonomously without questions?
│   └── Use: autonomous-skill-orchestrator
├── Multi-agent parallel work?
│   └── Use: agent-swarm-workflow
├── Creating/updating a skill?
│   └── Use: skill-creator
├── Memory issues or cleanup needed?
│   └── Use: memory-hygiene
└── Task tracking with beads?
    └── Use: beads-workflow

Auto-Discovery

This skill auto-discovers new skills added after its creation. Run the scan script to update the catalog:

bash skills/use-skill/scripts/scan-skills.sh

Important Notes

1. Always check this skill first - New skills may have been added since last session
2. Skills can be combined - Example: Use council to decide architecture, then autonomous-skill-orchestrator to implement
3. Some skills need setup - Check each skill's Prerequisites section before first use
4. Ping the user - Always announce when activating a skill: "Using skill: [name]"

Skills to Avoid (Redundant/Broken)

Skill	Reason	Alternative
god-mode	CLI not installable, docs incomplete	Manual git/gh commands
agentlens (npm)	No CLI binary, guidance only	Use codebase-cartographer (Rust CLI)
chromadb-memory	Redundant - we use LanceDB	Use memory-hygiene
supermemory	Requires OpenAI API + external DB	Use memory-hygiene

References

• Full skill details: See individual SKILL.md files in skills/[name]/
• New skills: Check skills/use-skill/scripts/scan-skills.sh output