finishing-a-development-branch
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes standard development tools including git, npm, cargo, pytest, and the GitHub CLI (gh) within SKILL.md. These commands are appropriately scoped to the skill's intended purpose of managing the lifecycle of development branches.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from test failure logs and git commit lists (SKILL.md). Boundary markers are provided by structured prompts that separate tool output from decision logic. The capability inventory includes local command execution and GitHub PR creation. While explicit sanitization of test output is absent, the rigid workflow structure mitigates the risk of an agent following embedded instructions.\n- [SAFE]: The skill implements significant safety controls, such as mandatory test verification before integration and requiring explicit typed confirmation ('discard') before performing destructive branch deletions.
Audit Metadata