#292292nd of 389 in Skills, plugins and prompts
yao-meta-skill
Agent skill that creates, evaluates, packages and governs other agent skills
- Stars
- 2.7k
- License
- MIT
- Last commit
- Nov 2026
- Language
- Python
285 on 21 Apr 20262,732 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
yao-meta-skill is a skill for Codex, Claude and other agents that turns workflows, transcripts and prompts into reusable skill packages. It compiles one platform-neutral Skill IR into OpenAI, Claude, generic Agent Skills and VS Code targets, and adds trigger and output evals, an HTML review page, package verification and release gates. A Python CLI (scripts/yao.py) drives these steps.
Who it is for: Teams building and maintaining reusable agent skills
Strengths
- Compiles one Skill IR into OpenAI, Claude, generic, Agent Skills and VS Code targets
- Output evals with blind A/B packs, answer keys and reproducible benchmark manifests
- Release checks: package verification, install simulation, permission probes, claim guard
- Telemetry is metadata-only, and durable writes need approval
Weaknesses
- Large surface of reports, gates and ledgers; heavy for a single simple skill
- Benchmark score is self-assessed; human review was one reviewer, five cases
- Provider-backed evidence, native permission execution and real-client telemetry still pending
- No tagged release; WorkBuddy import and routing are left as client acceptance checks
What it needs
- no GPU
- Needs Python 3, Node.js (npx skills CLI), GitHub (update check, benchmark scan)
- Models: OpenAI Codex, Claude, Agent Skills compatible agents, VS Code, WorkBuddy
Quick answers
Is yao-meta-skill open source?
Yes. Its code is on GitHub under the MIT license.
Also in Skills, plugins and prompts
See all 389| Rank | Project | Score |
|---|---|---|
| 1 | claude-memPersistent cross-session memory plugin for Claude Code and other coding agents | 97 out of 100 |
| 2 | impeccableDesign skill with 24 slash commands for AI coding agents' frontend output | 97 out of 100 |
| 3 | gstackMarkdown slash-command skills that give Claude Code planning, review, QA and ship roles | 96 out of 100 |
| #4 | rufloAgent harness adding swarms, memory and hooks to Claude Code and Codex | 96 out of 100 |
| #5 | cavemanShortens coding-agent replies and compresses tool output before the model reads it | 95 out of 100 |