The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill utilizes npx to execute the skillkit CLI tool directly from the NPM registry. It further provides commands to install and execute code from remote GitHub repositories.
[EXTERNAL_DOWNLOADS]: Fetches executable skills and configurations from various external sources. These include official repositories from trusted organizations (Anthropic, Vercel, Supabase, Stripe) and various community-contributed repositories.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and displays metadata from a marketplace containing over 400,000 third-party skills. This untrusted content could contain hidden instructions designed to manipulate the agent during the discovery process.
Ingestion points: External search results from npx skillkit find and marketplace browsing commands.
Boundary markers: None; the instructions do not define delimiters or specific warnings to ignore instructions embedded in search results.
Capability inventory: The skill possesses shell execution capabilities (npx) and the ability to perform file system writes during the installation of discovered skills.
Sanitization: There is no mention of sanitization, filtering, or validation of the external skill descriptions before they are processed by the agent.

find-skills