yaojingang/yao-meta-skill

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

What it solves

It transforms messy operational inputs—such as workflow notes, transcripts, prompts, and runbooks—into governed, reusable AI agent skills. It moves beyond simple prompt generation to create a full lifecycle system for modeling, compiling, evaluating, and maintaining AI assets that can be deployed across different platforms.

How it works

The project uses a "Skill OS" approach to manage the lifecycle of an agent skill:

  1. Intent Modeling: It starts with a human dialogue to define the job, outputs, and constraints explicitly.
  2. Skill IR (Intermediate Representation): It creates a platform-neutral semantic contract that separates the skill's intent from the specific platform format.
  3. Compilation: Target compilers and adapters translate the IR into formats compatible with OpenAI, Claude, VS Code, and other agent frameworks.
  4. Evaluation & Review: An "Eval Lab" and "Review Studio" generate evidence (assertions, timing, token usage) and gate reports to ensure the skill meets quality and trust standards before release.
  5. SkillOps Loop: It tracks adoption drift and telemetry to suggest adaptive proposals for the next iteration of the skill.

Who it’s for

  • AI Engineers and Operators who need to turn repeated manual workflows into standardized team assets.
  • Teams requiring strict governance, release gates, and evidence-based quality assurance for their AI agents.
  • Developers who want to build AI skills once and deploy them across multiple LLM providers or platforms.

Highlights

  • Cross-Platform Portability: Compiles a single skill model into multiple target adapters (OpenAI, Claude, etc.).
  • Evidence-Based Governance: Uses a "World-Class Evidence Ledger" and "Review Studio" to block premature claims and ensure release readiness.
  • Automated Benchmarking: Performs silent GitHub benchmark scans and reference synthesis to align skills with top public patterns.
  • Comprehensive Eval Lab: Provides blind A/B review packs, assertion grading, and runtime permission probes.
  • SkillOps Telemetry: Implements metadata-only adoption tracking to identify when a skill needs patching or updating.

Related

  • Project
    zjunlp/SkillNet

    SkillNet is an open‑source infrastructure that treats AI agent capabilities as reusable, searchable software assets. It offers a public skill library, a Python SDK/CLI, and a public REST API for discovering, downloading, creating, evaluating, analyzing, and orchestrating these skills. The platform enables agents to reuse existing functionality (e.g., PDF parsing, web‑scraping) instead of reinventing it, provides quality scores, builds skill‑relationship graphs, and can generate scene‑specific prompts for downstream LLM agents. Installation is as simple as `pip install skillnet-ai`; no API key is needed for search or public downloads, while creation/evaluation/orchestration require an LLM API key. The project includes a web explorer, MCP server, and integrations with Claude, OpenAI, OpenClaw, and JiuwenClaw, plus benchmark scripts for ALFWorld, WebShop, and ScienceWorld.

  • Project
    joeseesun/qiaomu-meta-skill

    A meta-skill that transforms prompts and workflows into researched, evaluated, and securely published AI agent skill packages.

  • Project
    HKUDS/OpenSpace

    A skill management layer for AI agents that enables them to retrieve, evaluate, and evolve their capabilities based on real-world task outcomes.

  • Project
    tripleyak/SkillForge

    An evidence-driven framework for creating and maintaining AI agent skills that uses behavioral testing to ensure every skill demonstrably improves performance.

  • Project
    chujianyun/skills

    A library of reusable Agent Skills and workflows for AI agents, covering everything from prompt optimization and document interpretation to system operations and career coaching.