load-issues

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the native system gh CLI tool via subprocess execution to retrieve issue details (gh issue list, gh issue view). It also invokes shell utilities such as mkdir -p to initialize target directories.\n- [INDIRECT_PROMPT_INJECTION]: The skill represents an indirect prompt injection surface because it reads external, untrusted content (GitHub issue titles, bodies, and author metadata) and dynamically formats it into local markdown files.\n
  • Ingestion points: External issue data returned from the gh issue view json schema payload is embedded dynamically into file structures.\n
  • Boundary markers: Absent. There are no delimiters or explicitly defined instructions forcing the agent to treat the <body> placeholder string as plain data or text formatting only.\n
  • Capability inventory: Local directory creation (mkdir -p) and file generation capability.\n
  • Sanitization: Absent. The instructions do not define any sanitization, escaping, or strict text processing rules prior to writing the <body> variable data block into filesystem markdown documents.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 01:09 AM
Security Audit — agent-trust-hub — load-issues