oma-scm

Pass

Audited by Gen Agent Trust Hub on Oct 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted content from Git repositories and has the authority to perform write operations like committing and pushing code.
  • Ingestion points: The skill reads repository history (git log), file differences (git diff), and project metadata files such as package.json, CHANGELOG, and CODEOWNERS across several resource files (SKILL.md, resources/cm-operations.md, resources/onboarding-risk-signals.md).
  • Boundary markers: No specific delimiters or "ignore instructions" guardrails are implemented to wrap the external data ingested from the repository to prevent it from being interpreted as agent instructions.
  • Capability inventory: The skill can execute powerful SCM commands including git add, git commit, git push, git merge, git rebase, and git worktree (SKILL.md).
  • Sanitization: While the skill correctly identifies and attempts to prevent the staging of secrets using tools like gitleaks and trufflehog, it does not provide mechanisms to filter or sanitize potential prompt injection attacks embedded in commit messages or code diffs.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 6, 2026, 04:52 PM