Deep Skill Finder
English | 中文
What is this?
Deep-skill-finder search powered by real execution and validation, not just follow creator claims
deep-skill-finder crawls Skills across the web, builds an evaluable candidate pool, and uses multi-channel recall to precisely match the best-fit Skill for your current task. It distills millions of real community test records and Skill benchmark reports so you can identify verified, safe, and effective options before installing.
Quick Start
Prerequisites
- A running Agent
Install deep-skill-finder in your Agent
- Send the following prompt directly to your Agent to complete installation:
Please install the deep-skill-finder skill: download the skill package from https://www.meyo.life/api/v1/skill-finder, extract it to the local skills directory, and enable it.
Use
Talk to your Agent in natural language — no special syntax needed. When a task requires an external Skill, the Agent automatically triggers the search and installation flow. Examples:
"Find me a skill that builds interactive dashboards from a CSV"
"Is there a skill for pulling stock market data?"
"Recommend a skill for translating technical docs into plain English"
"I need to set up a CI/CD pipeline for my repo, find me a skill that writes GitHub Actions configs that actually run"
"I'm building an AI product and need a business plan, find me a skill that helps structure the logic and polish the deck"
"I want to batch-transcribe audio interviews into clean text, find me a skill that does this end to end"
The Agent returns a ranked list of up to 5 recommendations with reasons. Confirm a number and the installation completes automatically.
Why not just search the marketplace?
Standard marketplace search has two fundamental problems:
-
Discovery — creators write broad, abstract descriptions to rank in more searches. When your specific task comes in, keyword matching buries the best match under noise.
-
Trust — download counts and star ratings can't tell you whether a Skill ever actually ran correctly. Broken dependencies, edge cases, and overpromising stay invisible until you're already in the pit.
deep-skill-finder addresses both: it recalls by actual task behavior, and ranks by real community run results — not by what a creator claims their Skill does.
Highlights
Web-wide Skill indexing — crawls Skill content across the web and builds an evaluable candidate pool, independent of any single platform's popularity ranking.
Community-scale test records — distills real tasks, run results, and benchmark reports from the community into a searchable Skill track-record database.
Multi-channel recall — pulls simultaneously from Skill capability descriptions, community test posts, and real outputs, preventing any single signal from missing the best candidate.
Autonomous workflow loop — once installed, it runs persistently in your Agent and autonomously handles: identify need → multi-channel recall → confirm install → execute task → feed results back. Each match gets more accurate over time.
Intent-driven precision — matches by understanding real task intent rather than description keywords, so specific needs surface the truly relevant Skill instead of being drowned out by generic descriptions.
Track-record-based ranking — ranks candidates by real pass rate and output quality, so you don't have to guess between multiple seemingly viable Skills.
Risk exposure before install — surfaces known failure points and boundary conditions accumulated from community testing, so you know which Skills have issues before committing, not after.
How it works
You describe a task in natural language
│
▼
Intent understanding
(rewrite into semantic query)
│
▼
Multi-channel recall
┌─────────┬──────────────┐
│ Skill │ Community │
│ profile │ test posts │
└────┬────┴──────┬───────┘
└─────┬─────┘
▼
Rank by verified track record
→ return TOP 5 with reasons
│
▼
Confirm a number → auto-install
Project structure
├── SKILL.md # Skill definition (read by the Agent)
└── scripts/
├── deep_skill_search.py # Semantic search via Meyo retrieval service
└── deep_skill_install.py # Download and install a Skill locally
Scripts reference
You normally don't call these directly — the Agent handles invocation. But you can run them standalone:
Search
python3 scripts/deep_skill_search.py "your task description" [--agent-type openclaw]
Install / Uninstall / List
# Install
python3 scripts/deep_skill_install.py <skill-name> --dir ~/.catpaw/skills
# Uninstall
python3 scripts/deep_skill_install.py <skill-name> --dir ~/.catpaw/skills --uninstall
# List installed
python3 scripts/deep_skill_install.py --dir ~/.catpaw/skills --list
Contributing
Issues and pull requests are welcome.
If you find that a specific Skill is ranked too high or too low, the underlying signal lives in the Meyo community — leaving a real test report there is the most direct way to improve future recommendations.
License
Released under the MIT License. Free to use, modify, and distribute — attribution appreciated.