tracer-finish-branch
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using
gitandgh(GitHub CLI) for branch management and cleanup. It also requires the execution of a project-specific<test command>to verify the implementation, which involves running arbitrary code within the local repository environment. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the local development environment that may be influenced by external project metadata or branch names.
- Ingestion points: In
branch-finishing.md, the skill ingests output from git commands, including directory paths, branch names, and merge bases (Step 2). - Boundary markers: The instructions lack explicit boundary markers or delimiters to differentiate between system instructions and data contained within branch names or test outputs.
- Capability inventory: The skill performs shell command execution (git, gh, test suites), file system modifications (merging branches, deleting branches), and workspace management (worktree removal) as detailed in
branch-finishing.md(Steps 4 and 5). - Sanitization: No sanitization or validation logic is present for variables such as
<feature-branch>or<base-branch>before they are interpolated into executable shell commands.
Audit Metadata