finishing-a-development-branch

Pass

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Potential surface for indirect prompt injection through the ingestion of untrusted repository data.
  • Ingestion points: The skill reads repository history (git log) and code changes (git diff) during the execution flow in SKILL.md.
  • Boundary markers: There are no explicit delimiters or instruction-isolation markers mentioned when this data is passed to sub-skills like the code reviewer.
  • Capability inventory: The agent has the ability to stage files, create commits, push to remote repositories, and generate GitHub Pull Requests.
  • Sanitization: No sanitization or filtering of the content within the git logs or diffs is performed before processing.
  • [COMMAND_EXECUTION]: Orchestrates branch management through the execution of system CLI tools.
  • Evidence: The skill utilizes git status, git commit, git push, and gh pr create as documented in the granular procedure steps in SKILL.md.
  • Mitigation: Security is maintained through a non-autonomous design that requires the agent to stop and wait for explicit user approval (ASK) before every state-changing action.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 17, 2026, 05:06 PM
Security Audit — agent-trust-hub — finishing-a-development-branch