finishing-a-development-branch
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local development tools and version control commands based on the project structure.
- It identifies and runs test runners like
npm,cargo,pytest,go test, andmakebased on the presence of configuration files. - It utilizes the GitHub CLI (
gh) for pull request creation and standardgitcommands for branch management and merging. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted local data to drive its logic, creating a potential injection surface.
- Ingestion points: Reads project configuration files (
package.json,Cargo.toml, etc.) and Git commit logs to determine execution commands and PR metadata. - Boundary markers: Uses
EOFheredocs to delimit the pull request body, which helps prevent unintended shell expansion of the content. - Capability inventory: Performs file system reads, local command execution (test runners, git), and network operations via the
ghandgitCLIs. - Sanitization: The skill relies on the underlying shell and CLI tools for basic sanitization, though the intent is to automate trusted local developer workflows.
Audit Metadata