How anbeime/skill works
This repository is primarily a catalog and updater for AI-agent skills, with a large collection of local skill packages. The main workflow is to run an update (once or on a schedule), collect and organize skill metadata, persist the catalog, then let users browse or export it. The README also documents catalog statistics and a skill-format validator. Some runtime wiring is inferred from the README and imports; unsampled implementation details are not represented as definite calls.
Main entry points: data_manager.py Data manager · main.py CLI entry point · scheduler.py Update scheduler · crawler.py Skill crawler
Components 18
Catalog update
- CLI entry point
main.py - Update scheduler
scheduler.py - Skill crawler
crawler.py - Update configuration
config.py
Catalog data
- Data manager
data_manager.py - Catalog JSON
data/skills.json - Local skill data
data/local_skills.json - CSV export
data_manager.py
Skill packages
- Skill packages
skills - Content skills
skills/content-creation-publisher - Video skills
skills/video-creation-suite - Document skills
skills/legal-assistant-skills-main - Agent teamwork
antinet-agentteams
Skill validation
- Skill validator
tools/skill_validator - Skill properties
tools/skill_validator/models.py
Other components
- Operator
- Skill source
- Skill store browser
Connections 14
- Operator → CLI entry point: runs commands
- CLI entry point → Update scheduler: starts updates
- CLI entry point → Data manager: reads or exports
- Update scheduler → Skill crawler: requests crawl
- Update scheduler → Data manager: persists results
- Skill crawler → Skill source: fetches skills
- Skill crawler → Update configuration: uses settings
- Data manager → Catalog JSON: reads and writes
- Data manager → Local skill data: reads local data
- Data manager → CSV export: exports CSV
- Skill store browser → Catalog JSON: browses catalog
- Skill store browser → Skill packages: browses skills
- Operator → Skill validator: checks skill
- Skill validator → Skill properties: checks properties