commit-push-pr

Pass

Audited by Gen Agent Trust Hub on Sep 22, 2026

Risk Level: SAFEDYNAMIC_CONTEXT_INJECTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DYNAMIC_CONTEXT_INJECTION]: The skill uses dynamic context injection (the ! syntax) to gather environment information at load time.
  • Evidence: !git branch --show-current, !git status --short, !git log --oneline -5, and !git diff --stat in SKILL.md.
  • The commands are read-only and restricted to git metadata relevant to the skill's core functionality.
  • [COMMAND_EXECUTION]: The workflow specifies execution of several CLI tools to perform its tasks.
  • Evidence: Commands include git add, git commit, git push, gh pr create, ruff format, ruff check, and pytest.
  • These are standard tools for a developer environment and align with the skill's description.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the local environment which could contain malicious instructions.
  • Ingestion points: Output from git status, git log, and git diff is loaded into the agent's context (SKILL.md).
  • Boundary markers: None present in the pre-computed context section.
  • Capability inventory: The skill can execute local commands (git, ruff, pytest, gh) and perform network operations (git push, gh pr create).
  • Sanitization: None visible; however, the skill explicitly instructs the agent to "Review Changes" and ensure no sensitive files are staged, which serves as a manual checkpoint.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 22, 2026, 07:39 PM
Security Audit — agent-trust-hub — commit-push-pr