# SkillPass

what it does

ai skills are starting to look like real software artifacts: instructions, workflow rules, permissions, tool assumptions, scripts, and setup requirements packaged for agents like Codex, Claude Code, Cursor, and similar tools.

SkillPass is the public surface for finding, inspecting, and installing those skills:

• browse skills by tool, category, risk, status, and maintainer
• keyword and AI search on the web and through the skillpass cli
• submit a github url or zip into a validation flow
• publish workflow packs as one listing instead of a pile of loose skills
• validate the package and show the findings before installation
• show a skill passport with schema, permissions, findings, risk, and source
• install, list, update, and remove local skills with receipts
• publish warnings clearly without pretending a skill is risk-free

the directory is live with a growing catalog, including workflow packs like AI Blueprint, single-skill listings, public passports, docs, sitemap metadata, and maintainer controls.

You can browse official sources and public maker pages, then use a tool-specific install guide. The CLI also supports installing from a GitHub repository without requiring it to be listed in the directory first.

For authors, skillpass init scaffolds a SKILL.md and manifest, asks about permissions, and scans the result. For a one-off task, skillpass use prepares a listed skill in a temporary directory and outputs its prompt after preflight and hash verification. It does not permanently install the skill or launch the agent itself. Packs and direct GitHub references still use the install path.

why i built it

the useful part of an agent skill is also the risky part. a skill can ask the agent to read files, run shell commands, use connectors, send data elsewhere, or make changes in a repo. that should not be copied around as random markdown with no inspection layer.

i wanted a place where skills can be shared, but the default posture is still skeptical. every skill starts untrusted. SkillPass earns confidence by showing what was checked, what was declared, what was detected, and what still needs human judgment. the cli keeps that posture on the install side too: preflight checks, permission visibility, hash verification, and local receipts.

how it’s built

a quick tour:

• astro + react islands: fast public pages with interactive submission and validation surfaces
• astro ssr: dynamic skill, version, and maintainer pages stay fresh after publish
• hono api: github sign-in, authenticated submissions, ownership rules, and publish routes
• drizzle + neon: users, submissions, validation jobs, versions, and public listing data
• cloudflare r2: content-addressed snapshots and uploaded zip archives
• skill schema package: shared zod contracts for manifests, reports, passports, submission enums, and wire types
• validator package: checks structure, secrets, risky instructions, and undeclared permissions
• node cli: discovery, local validation, and receipt-backed installation, plus author scaffolding and one-off use of listed skills
• maintainer dashboard: unlist and relist owned skills while keeping versions, snapshots, and passports intact

what’s next

The directory and CLI are live. The work now is hardening the package path, especially binary-file handling, and keeping validation limits visible. A passport is an inspection report, not a guarantee that a skill is safe.

discovery is useful only if the install step gets more skeptical, not less.