git
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides workflow refinements for using Git, such as branch naming conventions (avoiding prefixes) and Pull Request merge preferences using the GitHub CLI (
gh). - [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest user-provided GitHub issue numbers to be included in commit messages. While this creates a surface where untrusted data is interpolated into shell commands, the instructions explicitly acknowledge the risk of shell command substitution (specifically regarding backticks) and provide safe remediation strategies, such as using single-quoted here-docs or escaping backticks to prevent unintended execution.
Audit Metadata