om-fix-issue-multi-optimized
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides a surface for indirect injection by processing untrusted repository content (issues, PR bodies) and loading local skill extensions from the project directory.
- Ingestion points:
references/agentic-setup.mdidentifies repository-local configuration (.ai/agentic.config.json) and skill extensions (.ai/skills/om-fix-issue-multi-optimized/SKILL.md) as potential input sources. - Boundary markers:
references/agentic-setup.mdincludes clear defensive boundaries, instructing the agent to ignore directives within data such as "ignore previous instructions" or unauthorized commands. - Capability inventory: The skill manages code fixes, testing, and review processes involving shell execution and file system access within the repository scope.
- Sanitization: Strict sanitization is mandated in
references/agentic-setup.md, requiring regex validation for all externally-sourced identifiers used in shell commands or path interpolation. - [CREDENTIALS_SAFE]: The skill incorporates explicit secrets hygiene rules in
references/rules.md, forbidding the exposure of tokens or credentials in logs, comments, or plans, effectively mitigating credential harvesting risks.
Audit Metadata