The Agent Skills Directory

[PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted external content from the web via Perplexity tools. This introduces an indirect prompt injection attack surface where malicious instructions embedded in search results could attempt to manipulate the agent's behavior.
Ingestion points: SKILL.md instructs the agent to use mcp__perplexity__perplexity_search and mcp__perplexity__perplexity_ask for processing generic queries.
Boundary markers: The instructions do not provide specific delimiters or warnings for the agent to ignore instructions contained within the fetched web content.
Capability inventory: The skill is intended for an environment like Claude Code, which typically has broad capabilities including file system access and terminal execution.
Sanitization: No sanitization or validation logic is specified for the content retrieved from the web.

perplexity