The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data from Linktree's documentation and status pages, as well as local repository files, creating a surface for indirect prompt injection.
Ingestion points: The skill uses WebFetch to ingest content from help.linktr.ee and status.linktr.ee. It also uses Read, Glob, and Grep to inspect local files such as references/official-docs.md and campaign specifications.
Boundary markers: The instructions include clear boundaries, requiring the agent to "separate observed facts from assumptions" and use "redacted receipts."
Capability inventory: The skill utilizes Write and Edit tools for recording results and WebFetch for network operations.
Sanitization: The instructions explicitly forbid searching credential files and mandate the redaction of sensitive information from receipts.

linktree-hello-world