github-issues

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill includes logic to automatically retrieve GitHub authentication tokens from local configuration files.\n
  • Evidence: Shell scripts in SKILL.md parse ~/.hermes/.env and ~/.git-credentials to extract secret tokens.\n- [COMMAND_EXECUTION]: The skill utilizes inline Python execution to process data received from external API calls.\n
  • Evidence: Frequent use of python3 -c to parse and format JSON data returned by curl requests.\n- [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection from external user-generated content on GitHub.\n
  • Ingestion points: Fetches issue titles, bodies, and labels from api.github.com (SKILL.md).\n
  • Boundary markers: No delimiters or explicit instructions are provided to the agent to ignore embedded instructions in the fetched data.\n
  • Capability inventory: Execution of shell commands and python snippets via subprocesses.\n
  • Sanitization: None; external data is directly formatted and presented to the agent context.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 04:32 PM
Security Audit — agent-trust-hub — github-issues