Chat context vanishes. Markdown skills drift. Sealed .skill packages carry typed I/O, workflow, and inspect-before-run trust — install once, then talk to your AI.
.skill packages The Open .skill Protocol is an open specification and portable .skill format so AI agents can author, inspect, hand off, and run skills across hosts. skillerr is the reference implementation CLI.
Format: .skill (sealed ZIP) · Status: 1.0.0 (Stable)
Reference CLI: skillerr (bin: skill)
Repo: dot-skill/skillerr
npm i -g skillerrNode ≥ 20. After that, talk to your agent — not a human CLI checklist.
Create from this chat
Run these exact commands in your terminal, in order:
1. npm i -g skillerr (skip if `skill --version` already works)
2. export SKILL_HOST=cursor (replace "cursor" with your actual tool name)
Then, from this conversation, create a portable .skill with a redacted journey
and exact sections I approved (secrets as {{refs}}). Checkpoint for handoff, or
compile --approve --mint when release-complete. Do not invent filler. Show
status and the output path.Inspect before run
Inspect ./file.skill TrustView without executing. Validate, then dry-run.
Summarize trust warnings. Do not execute for real unless I ask.More starters: Getting started · Convert a SKILL.md · Evaluate & score · Agents · Workflows
.skill example.skill
├── skill.json # manifest, digests, profile, completeness
├── workflow.json # runnable steps
├── knowledge/ # pinned decisions / rules
├── provenance/ # redacted journey + generation_usage
└── signatures/ # mint attestation (release)| Surface | Role |
|---|---|
skillerr | Reference CLI |
@skillerr/protocol | SkillContract, schemas, types |
@skillerr/core | Compile, pack, validate, mint |
@skillerr/runtime | Inspect / dry-run / execute |
@skillerr/workspace | Local .skill/ working tree |
@skillerr/registry | Optional local transparency log |