oh-merge

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes git and gh commands to manage branches and merge PRs. These are standard operations for development workflows. Examples include gh pr list, gh pr merge, git rebase, and git push --force-with-lease.
  • [PROMPT_INJECTION]: The skill processes untrusted data such as PR titles, descriptions, and code diffs. This creates a potential surface for indirect prompt injection attacks where instructions could be embedded in the code or PR metadata. However, the risk is mitigated by the requirement for a human-applied oh-merge label and mandatory completion of the /ship pipeline, which requires manual verification and specific approval comments.
  • Ingestion points: PR metadata and file diffs via GitHub CLI.
  • Boundary markers: Relies on external ship/review gates and the presence of specific human-applied labels.
  • Capability inventory: Includes branch manipulation (git rebase, git push --force-with-lease), PR state changes (gh pr edit), and merging (gh pr merge).
  • Sanitization: Relies on separate review sub-agents and human verification prior to the merge process.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 04:54 PM
Security Audit — agent-trust-hub — oh-merge