This is where it started
“The more skills your AI agent can use, the more you have to re-check, every single time, whether it's actually safe to use one.”observed → Started 2026.07.03
— So we built
SkillBoard
Livetoolsnpm package
A CLI that connects AI-agent skill selection to project-scoped policy.
Developer evidence
What I owned in this project
- Period
- 2026.07 – Operating now
- Team
- Solo developer
- Contribution
- Planning · design · development · deployment · operations, 100%
- Owned scope
- Combines skill inventory, workflow policy, guard state, action cards, and bridge cleanup into one CLI surface.
Introduction
SkillBoard shows AI agents which skills they may use in the current project and why others are blocked. It separates installed skills from allowed ones and keeps policy changes behind an explicit approval step.
User flow
- 01
Initialize SkillBoard in a project and scan skill sources and workflows.
- 02
Use brief and route to see usable skills and why others are blocked.
- 03
Approve policy changes through action cards, then dry-run cleanup before removal when it is no longer needed.
Highlights
- Summarizes project-scoped skill policy and current availability through a short brief.
- Checks project policy and safety conditions before a skill is used.
- Keeps policy changes explicit and reviewable.