Skills
Navigate
skills/undermybelt/hermes-skills/offensive-osint/Gen Agent Trust Hub

offensive-osint

Fail

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: CRITICALEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The scripts/h1_reference.py script fetches public vulnerability disclosures from HackerOne's GraphQL API (hackerone.com). HackerOne is a well-known security platform, and the interaction is restricted to public data via standard network requests.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data from HackerOne reports which could theoretically contain malicious instructions designed to influence an AI agent.
  • Ingestion points: External data is ingested via the HackerOne GraphQL API in scripts/h1_reference.py.
  • Boundary markers: The output from the HackerOne API is printed to stdout without delimiting characters or instructions to the agent to ignore embedded commands.
  • Capability inventory: The skill includes tools for network communication (scripts/h1_reference.py) and local file system scanning (scripts/secret_scan.py).
  • Sanitization: The script performs no sanitization or escaping of the report titles or summaries before they are returned to the agent context.
Recommendations
  • Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Sep 17, 2026, 08:33 PM