</>MCP Agents Market
Skill

yao-meta-skill

by yaojingang2.6kPythonUpdated 2026-08-17

YAO = Yielding AI Outcomes. A rigorous engineering, evaluation, governance, and portability system for reusable agent skills.

CodexClaudeOpenAIAgent SkillsVS Code

YAO Meta Skill (Yielding AI Outcomes) is a comprehensive agent skill engineering system that transforms workflows, prompts, and operational knowledge into reusable, governed skill packages for AI agents. It provides a complete Skill OS 2.0 framework with semantic contracts, target compilers for multiple platforms (OpenAI, Claude, Agent Skills, VS Code), output evaluation labs, and release governance. The system emphasizes rigorous testing with train/dev/holdout evaluation suites, benchmark reproducibility, and evidence-based quality gates to ensure skills are production-ready team assets rather than one-off prompts.

Key Features

Skill IR (Intermediate Representation) - platform-neutral semantic contract for intent, triggers, inputs, outputs, and boundaries
Multi-target compilation to OpenAI, Claude, generic agent skills, Agent Skills packages, and VS Code workflows
Output Eval Lab with assertion grading, execution evidence, timing/token tracking, blind A/B review packs, and adjudication reports
Review Studio 2.0 with HTML gate pages combining intent, eval results, context cost, runtime checks, trust signals, and release evidence
Governance system with ownership, lifecycle states, maturity scoring, promotion policies, and world-class evidence ledgers
Trigger evaluation using train/dev/holdout/adversarial suites with semantic intent models and route confusion tracking
Metadata-only adoption drift telemetry, SkillOps reports, and adaptive proposal loops for continuous improvement
Package verification with install simulation, runtime permission probes, evidence consistency checks, and claim guards

Use Cases

  • 01Convert repeated workflows or prompt sets into installable, cross-platform skill packages with explicit contracts
  • 02Upgrade personal skills into team assets with interface contracts, output evals, and review evidence before wider adoption
  • 03Prepare skills for beta release with package verification, compatibility checks, and evidence consistency validation
  • 04Maintain released skills with metadata telemetry, adoption drift detection, and SkillOps reports to guide iteration
  • 05Build governed skill libraries with ownership, maturity scoring, review cadence, and portfolio-level drift signals
  • 06Compare meta-skill approaches using weighted quality benchmarks across method depth, eval rigor, and governance

Related Skills

View more

yao-meta-skill — FAQ

What is YAO Meta Skill?+

YAO Meta Skill is a rigorous engineering system for creating, evaluating, packaging, and governing reusable AI agent skills. It transforms workflows, prompts, and operational knowledge into portable skill packages with semantic contracts, multi-platform compilation, and evidence-based quality gates.

How do I install YAO Meta Skill?+

Install globally for Codex using npx: 'npx -y skills add yaojingang/yao-meta-skill -a codex -g -y'. For all supported agents, use 'npx -y skills add yaojingang/yao-meta-skill -a '*' -g -y'. Restart your AI client after installation.

Which AI clients does it work with?+

YAO Meta Skill compiles skills for OpenAI, Claude, generic agent platforms, Agent Skills-compatible systems, and VS Code/Copilot workflows. It can be installed as a skill in Codex and other agent systems that support the Agent Skills format.

What are the prerequisites?+

Python 3.11 or newer is required for local development and running the validation/compilation scripts. No API keys are needed; the system works locally and generates portable skill packages.

Is YAO Meta Skill free to use?+

Yes, it is open source under the MIT license and free to use. All features including skill creation, evaluation, compilation, and governance are available without cost.

How does it differ from other skill creators?+

YAO emphasizes engineering rigor with weighted quality benchmarks showing 91.5/100 vs Anthropic's 67.5 and OpenAI's 50.5. It provides complete eval/test suites, governance systems, and portability that other approaches lack, though it has steeper onboarding compared to conversational skill creators.

How do I install yao-meta-skill?+

Open the source repository on GitHub and follow its README. yao-meta-skill is a skill — MCP Agents Market links you directly to the official repo.

Is yao-meta-skill free?+

yao-meta-skill is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.

Related searches