firecrawl-debug-bundle
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEDATA_EXFILTRATIONDYNAMIC_EXECUTIONDYNAMIC_CONTEXT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill accesses sensitive local files including
.envand system logs in/tmp/to collect diagnostic information. It includes a redaction step usingsedto mask common secret patterns (API_KEY, SECRET, TOKEN, PASSWORD) before inclusion in the bundle. Users are instructed to review the final archive before sharing it with support. - [DYNAMIC_EXECUTION]: The skill generates and executes local shell and TypeScript scripts to perform API connectivity tests against official Firecrawl endpoints and package the diagnostic results.
- [DYNAMIC_CONTEXT_INJECTION]: Uses the
!command syntax to execute environment checks (node --versionandnpm list) at skill load time to verify prerequisites. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from system logs in
/tmp/. While it uses simple redaction, it lacks formal boundary markers or sanitization for the log content, though the impact is limited by the skill's diagnostic purpose.
Audit Metadata