The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources to verify claims, creating a surface for potential indirect prompt injection.
Ingestion points: The skill reads data from API responses via curl, package metadata through npm and pip, repository information using gh, and local file contents using ls and grep.
Boundary markers: There are no explicit boundary markers or delimiters mentioned to isolate external data from the agent's internal instruction context.
Capability inventory: The skill utilizes several powerful tools including file system access (ls, grep), network operations (curl), package management inspection (npm, pip), and version control tools (git, gh).
Sanitization: The skill demonstrates security awareness by instructing the agent to only check HTTP status codes (-o /dev/null) when verifying user-provided URLs to avoid fetching or rendering potentially malicious content.

Fact Check