branch-complete
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes standard source control commands and platform-specific CLI tools including
git,gh(GitHub),glab(GitLab), andaz(Azure DevOps) to automate branch management and Pull Request creation. - [COMMAND_EXECUTION]: Executes a placeholder
[test command]which is intended to be a project-specific verification script. This is standard behavior for development automation tools. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by reading content from local project artefacts (markdown files like stories and test plans) and interpolating them into PR descriptions. However, it mitigates potential shell injection risks by using industry-standard practices such as HEREDOCs with quoted delimiters (
'EOF') in bash and temporary files for body content in PowerShell. - [SAFE]: Manages project-specific state and health metadata in a local file (
.github/pipeline-state.json) within the repository scope. No sensitive files or credentials are accessed or exposed.
Audit Metadata