finishing-a-development-branch
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes several standard CLI tools such as
gitfor branch management,ghfor GitHub integration, and common test runners likenpm,cargo,pytest, andgo. These are used to automate the final stages of the development lifecycle. - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill processes local repository data including branch names, commit lists, and worktree paths, as well as user-supplied input for Pull Request titles and descriptions in
SKILL.md. - Boundary markers: It utilizes quoted heredocs (
'EOF') when constructing shell commands for PR creation, which prevents the content of the PR body from being interpreted as executable shell code. - Capability inventory: The skill has the capability to modify the local file system (git operations), communicate with remote repositories (
git push,gh pr create), and execute arbitrary test suites defined in the project. - Sanitization: Includes manual confirmation steps for destructive operations like discarding work and uses structured options to limit the agent's action space.
Audit Metadata