SkillAnything
Auto-generate production-ready agent skills for any CLI or workflow.
About
A meta-skill: a 7-phase pipeline (analyze, design, scaffold, test plan, evaluate, optimize, package) that turns a CLI tool, API, or workflow into a multi-platform skill for Claude Code, OpenClaw, and Codex. It even benchmarks the generated skill with and without itself and optimizes the trigger description via a train/test loop.
Generated skills are only as good as the target's documentation — poorly documented CLIs produce thin skills — and the full 7-phase run is heavyweight for simple cases where a hand-written SKILL.md would do.
What it can do
- CLI/API target analysis
- Skill architecture design
- SKILL.md + scripts + references scaffolding
- Eval-set generation and benchmarking
- Trigger-description optimization
- Multi-platform packaging
Setup
- Clone for Claude Code: git clone https://github.com/AgentSkillOS/SkillAnything.git ~/.claude/skills/skill-anything
- For OpenClaw: git clone https://github.com/AgentSkillOS/SkillAnything.git ~/.openclaw/skills/skill-anything
- For Codex: git clone https://github.com/AgentSkillOS/SkillAnything.git ~/.codex/skills/skill-anything
- Then ask, e.g. 'Create a skill for the httpie CLI tool'
Prerequisites
Python for the pipeline scripts; git
Use cases
Works with
Trust score — 5/10
Scored 5/10 (metrics tier, metrics-v1). Evidence: Meta-skill: 7-phase pipeline for generating new skills; Namespace discrepancy: repo page at tanapoljoe1-art/skillanything, install commands reference AgentSkillOS/SkillAnything — verify before cloning; Generated-skill quality bounded by target documentation quality; Full pipeline is heavyweight for simple cases. License: License not confirmed in sources read.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 5.0/10
- Maintenance · recency, releases, issue load
- 6.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 5.0/10
- · Meta-skill: 7-phase pipeline for generating new skills
- · Namespace discrepancy: repo page at tanapoljoe1-art/skillanything, install commands reference AgentSkillOS/SkillAnything — verify before cloning
- · Generated-skill quality bounded by target documentation quality
- · Full pipeline is heavyweight for simple cases
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.