occ-diagnose

Pass

Audited by Gen Agent Trust Hub on Oct 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes npx to run @ruvnet/open-claude-code@2.1.3 for version and help verification. This is a vendor-owned package and is central to the skill's diagnostic function.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to inspect skill definition files located in .claude/skills/ and ~/.claude/skills/. This introduces a surface for indirect injection from local configuration files; however, the skill includes explicit mandates to sanitize output and redact sensitive provider data.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 12, 2026, 04:50 PM