performing-ip-reputation-analysis-with-shodan

Pass

Audited by Gen Agent Trust Hub on Apr 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Provides a functional Python agent (scripts/agent.py) designed to be executed via CLI to perform IP lookups.
  • [EXTERNAL_DOWNLOADS]: Dependencies include standard, well-known Python packages shodan and requests for API interaction.
  • [DATA_EXFILTRATION]: Performs network requests to legitimate threat intelligence platforms (Shodan and AbuseIPDB) to retrieve IP metadata. These operations are transparent and essential to the skill's stated purpose.
  • [PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection via external data ingestion.
  • Ingestion points: Service banners and metadata retrieved from the Shodan API processed in scripts/agent.py.
  • Boundary markers: No specific delimiters are implemented to separate API-returned data from the agent's internal instruction context.
  • Capability inventory: The script is limited to network lookups and printing results; it does not possess autonomous file-writing or system-level execution capabilities.
  • Sanitization: External API content is processed without specific sanitization or validation beyond truncation in documentation examples.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 9, 2026, 06:48 PM