Motivation
The trend of "Skill Standardization" is gaining significant traction in the AI Agent ecosystem. Projects like mattpocock/skills and obra/superpowers are demonstrating that treating skills as modular, version-controlled code units ("Skill-as-Code") drastically improves reusability and collaboration. Meanwhile, Panniantong/Agent-Reach highlights the need for agents to easily discover and consume external data sources.
Currently, iflytek/skillhub excels as an enterprise registry with RBAC and audit logs, but it lacks a standardized ingestion mechanism for these emerging "Skill-as-Code" formats. Developers often have to manually map directory structures or write custom scripts to import skills from GitHub repos or local directories into the registry, which breaks the "local-first" and "git-centric" workflows popular in the open-source community.
Proposed feature
Add a skillhub import CLI command (or a dedicated API endpoint) that supports:
- Directory-to-Registry Import: Automatically parse a local directory or a git repository following a standard schema (e.g.,
manifest.yaml or skill.json at the root) and push the skill package to the registry.
- Semantic Versioning (SemVer) Support: Integrate with
semantic-release or similar tools to automatically bump versions (major/minor/patch) based on commit history or PR labels during the import/publish process.
- Pre-release Tagging: Support
alpha, beta, and rc tags to allow enterprises to test new skills in staging environments before promoting them to stable.
Why it fits this repo
iflytek/skillhub is positioned as the "Self-hosted, open-source agent skill registry for enterprises." To become the de-facto enterprise standard for skill governance, it must bridge the gap between open-source "Skill-as-Code" practices and enterprise-grade governance (RBAC, Audit). By supporting standard directory structures and SemVer, SkillHub can act as the central hub for both internal proprietary skills and curated external open-source skills, aligning with the "modular AI" trend.
References
Motivation
The trend of "Skill Standardization" is gaining significant traction in the AI Agent ecosystem. Projects like
mattpocock/skillsandobra/superpowersare demonstrating that treating skills as modular, version-controlled code units ("Skill-as-Code") drastically improves reusability and collaboration. Meanwhile,Panniantong/Agent-Reachhighlights the need for agents to easily discover and consume external data sources.Currently,
iflytek/skillhubexcels as an enterprise registry with RBAC and audit logs, but it lacks a standardized ingestion mechanism for these emerging "Skill-as-Code" formats. Developers often have to manually map directory structures or write custom scripts to import skills from GitHub repos or local directories into the registry, which breaks the "local-first" and "git-centric" workflows popular in the open-source community.Proposed feature
Add a
skillhub importCLI command (or a dedicated API endpoint) that supports:manifest.yamlorskill.jsonat the root) and push the skill package to the registry.semantic-releaseor similar tools to automatically bump versions (major/minor/patch) based on commit history or PR labels during the import/publish process.alpha,beta, andrctags to allow enterprises to test new skills in staging environments before promoting them to stable.Why it fits this repo
iflytek/skillhubis positioned as the "Self-hosted, open-source agent skill registry for enterprises." To become the de-facto enterprise standard for skill governance, it must bridge the gap between open-source "Skill-as-Code" practices and enterprise-grade governance (RBAC, Audit). By supporting standard directory structures and SemVer, SkillHub can act as the central hub for both internal proprietary skills and curated external open-source skills, aligning with the "modular AI" trend.References