git-init
Pass
Audited by Gen Agent Trust Hub on Feb 28, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill includes a 'CRITICAL' instruction requiring the agent to ignore any user-provided arguments or paths, which serves as a robust defense against direct prompt injection attempts.- [COMMAND_EXECUTION]: The skill explicitly prohibits the use of bash or shell commands for detection and file writing, instead mandating the use of safer, purpose-built tools like Glob and Write.- [DATA_EXFILTRATION]: No network operations or external data transfers are present. The skill only interacts with local project metadata (file names and .gitignore content) and explicitly forbids reading sensitive file contents.- [SAFE]: The skill employs a transparent workflow requiring user confirmation and previews, using static, standard industry templates for configuration.
Audit Metadata