performing-osint-with-spiderfoot
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data discovered during OSINT scans (such as subdomains, email addresses, and leaked credential content) and incorporates it into a structured report for the agent, which could contain malicious instructions designed to influence agent behavior.
- Ingestion points:
scripts/agent.pyretrieves scan results from the SpiderFoot REST API using theget_scan_resultsfunction. - Boundary markers: Absent. The script does not utilize delimiters or specific instructions (e.g., 'ignore content within this block') to protect the agent from embedded malicious instructions within the data.
- Capability inventory: The script performs network requests (GET/POST) to a SpiderFoot server and writes the resulting OSINT profile to a local JSON file.
- Sanitization: No sanitization, validation, or filtering is performed on the data elements retrieved from the external source before they are presented to the agent.
Audit Metadata