wheelry/deep-skill-finder

Deep Skill Finder – Real Tests Real Reviews find skills that Real Work

What it solves

It addresses the difficulty of finding the right AI agent skill for a specific task. While many skill directories exist, they often rely on popularity metrics like stars or downloads, which fail to indicate whether a skill actually works for a user's specific environment, task, or constraints.

How it works

It acts as an agentic discovery engine that integrates directly into your AI agent (such as Claude Code, Cursor, or Cline). When you describe a task, it performs a two-stage process: first, it retrieves relevant candidates based on your specific intent and environment; second, it ranks them using a database of real-world execution feedback—including success rates, performance in similar scenarios, and user-submitted reviews—rather than just popularity.

Who it’s for

It is designed for developers and power users who utilize AI agents for automation and need a reliable way to discover and verify third-party skills before installing them.

Highlights

  • Context-Aware Discovery: Matches skills based on your specific task, agent type, and environment requirements.
  • Real-Scenario Feedback: Ranks skills based on actual execution performance and user-provided reviews rather than generic popularity metrics.
  • Broad Compatibility: Works across 40+ agent runtimes, including Claude Code, Cursor, Windsurf, and Cline.
  • Privacy-Focused: Uses deterministic redaction to ensure that raw task trajectories are never uploaded when sharing feedback.
  • Simple Integration: Installs as a skill in your agent in under 30 seconds via a single prompt.

Related

  • Project
    catlog22/maestro-flow

    Maestro‑Flow is a TypeScript/Node.js framework that turns natural‑language intents into full‑stack development pipelines executed by multiple LLM agents (Claude, Gemini, Codex, etc.). Its Ralph engine classifies intents into 40+ command chains, dynamically decides at “decision” checkpoints whether to continue, roll back, or insert repair loops, and stores discovered patterns in a SQLite knowledge graph that is automatically injected into later runs. It supports four coordination modes (Delegate, Team, Wave, Swarm), long‑running autonomous loops via the Odyssey command, and precise literal/semantic code search. Installation is via `npm i -g maestro-flow`, and the tool provides extensive CLI and web‑dashboard documentation.

  • Project
    TsinghuaC3I/MARTI

    An open-source framework for training LLM-based multi-agent systems using reinforcement learning, featuring tree search augmentation for complex reasoning and code generation.

  • Project
    Dwinovo/minecraft-numen

    Numen is an open‑source Minecraft mod that embeds an embodied AI agent controllable by large language models. Users install the mod, configure an OpenAI/Anthropic API key, summon a bot, and interact with it via natural language, voice, or scripted Lua actions. The project provides a three‑layer architecture (Minecraft core, Lua action API, LLM sandbox) and a plugin system for extending the agent’s capabilities.

  • Project
    FoundationAgents/ReCode

    ReCode is an LLM agent framework that uses recursive code generation to unify high-level planning and low-level action, enabling dynamic granularity control for complex tasks.