branch-api-diff-design
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill instructions define strict output formats and technical priorities without attempting to override the agent's safety guardrails or bypass system instructions.
- [DATA_EXFILTRATION]: Analysis is limited to local Git repository metadata and source code files. No network transmission logic or unauthorized access to sensitive file paths was observed.
- [COMMAND_EXECUTION]: The skill utilizes the git tool for legitimate fetch and diff operations. No unsafe command construction or arbitrary shell execution patterns were found.
- [SAFE]: The skill follows best practices for its intended technical purpose and does not incorporate external dependencies or remote code. All identified behaviors are consistent with a benign developer tool.
Audit Metadata