firecrawl-search
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes shell commands via
firecrawlandnpx firecrawl-clito perform web searches, scrapers, and data tool discovery. - [EXTERNAL_DOWNLOADS]: Utilizes
npxto download and execute thefirecrawl-clipackage from the public npm registry at runtime. - [DATA_EXFILTRATION]: Sends usage telemetry, including search IDs, query ratings, and lists of valuable or missing sources, to the service backend using the
search-feedbackcommand. - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill fetches and ingests untrusted web content and search excerpts directly into the agent's context (SKILL.md).
- Boundary markers: There are no instructions providing delimiters or warnings to ignore embedded instructions within the scraped web data.
- Capability inventory: The skill has extensive shell access via the
firecrawlandnpxtoolsets. - Sanitization: No sanitization, escaping, or filtering of the external web content is performed before it is processed by the agent.
Audit Metadata