firecrawl-knowledge-ingest

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from external URLs (documentation portals). While this is the primary purpose, it introduces a surface where malicious instructions hidden on a website could potentially influence the agent.
  • Ingestion points: Content is scraped from external URLs and docs portals using the Firecrawl browser tool.
  • Boundary markers: The instructions specify stripping navigation chrome, headers, and footers, which provides a basic level of content isolation, though it does not explicitly warn against embedded instructions.
  • Capability inventory: The skill focuses on extraction and structured output (JSON/markdown) and does not call dangerous subprocesses or perform unauthorized file writes.
  • Sanitization: The skill focuses on preserving formatting like tables and code blocks without explicit sanitization steps mentioned.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill requires a FIRECRAWL_API_KEY. This is requested through a required input field in the frontmatter rather than being hardcoded, which follows best practices for secret management in agent skills.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 12:46 PM
Security Audit — agent-trust-hub — firecrawl-knowledge-ingest