bellingcat-image-video

Pass

Audited by Gen Agent Trust Hub on Jul 12, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No prompt injection attempts, safety bypass instructions, or adversarial framing patterns were detected. The instructions are purely functional and align with the stated purpose of identifying OSINT tools.
  • [DATA_EXFILTRATION]: No evidence of hardcoded credentials, sensitive file path access, or unauthorized network operations. The listed URLs point to legitimate investigative tools and documentation.
  • [EXTERNAL_DOWNLOADS]: The skill does not perform any automated downloads or package installations. It contains Markdown links to external websites for manual use or reference by the agent.
  • [REMOTE_CODE_EXECUTION]: There are no patterns involving the execution of remote scripts, piped shell commands, or dynamic code evaluation.
  • [COMMAND_EXECUTION]: The skill contains no shell commands or subprocess calls. It is an informational Markdown file.
  • [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or private secrets were found within the skill file.
  • [SAFE]: The skill serves as a reference guide for legitimate OSINT tools. References to well-known services (Google, Amazon, Microsoft) and established investigative organizations (Bellingcat, RAND Corporation) are documented neutrally and follow safe practices.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 12, 2026, 06:49 AM
Security Audit — agent-trust-hub — bellingcat-image-video