The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from the web using the mcp__perplexity-ask__perplexity_ask tool, creating a surface for indirect prompt injection attacks.
Ingestion points: External content retrieved from web search results via Perplexity.
Boundary markers: There are no explicit delimiters or instructions provided to help the agent distinguish between its own instructions and content from the web results.
Capability inventory: The skill includes file system tools (Read, Grep, Glob) that could be manipulated if the agent follows instructions hidden in search results.
Sanitization: The skill lacks instructions for sanitizing or validating the content returned from external sources before further processing.

researching-web