The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill enables the agent to ingest untrusted content from the internet through neural search results and page crawling, creating a surface for potential indirect prompt injection attacks.
Ingestion points: Data enters the agent's context through the output of tools defined in SKILL.md, specifically web_search_exa, web_search_advanced_exa, get_code_context_exa, and crawling_exa.
Boundary markers: No specific delimiters or safety instructions regarding external content are defined in the tool schemas.
Capability inventory: The skill provides tools for broad web search, advanced filtered search, code context retrieval, company/people research, and full-page crawling.
Sanitization: No explicit sanitization or filtering of external content is specified in the skill's instructions; the agent relies on its primary safety guardrails to handle untrusted data.
[EXTERNAL_DOWNLOADS]: The setup instructions in SKILL.md recommend configuring the environment using npx, which fetches the exa-mcp-server package from the official NPM registry. This is a standard and expected deployment method for official Exa integrations.

exa-search