ask-navigator

Pass

Audited by Gen Agent Trust Hub on Oct 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local script node scripts/shipyard-audit.mjs to perform repository audits.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external issue trackers (GitHub, GitLab), creating a vulnerability surface for indirect prompt injection.\n
  • Ingestion points: External issue tracker bodies and comments.\n
  • Boundary markers: Absent.\n
  • Capability inventory: File system writes to .omc/, docs/, and CONTEXT.md, tracker API interaction, and background subagent spawning.\n
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 13, 2026, 03:24 AM