wizard
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill determines setup requirements by reading local repository files such as
README.md,.env.example, and GitHub workflows. While malicious instructions in these files could theoretically attempt to influence the agent's proposed setup stages, the agent only generates a script for the human user to review and execute manually. - [COMMAND_EXECUTION]: The provided bash template includes helper functions that invoke the
gh(GitHub CLI) to set repository secrets and variables. This behavior is a documented feature of the skill intended for CI/CD setup and operates within the user's local security context. - [DATA_EXPOSURE]: The generated scripts are designed to collect and store sensitive information like API keys. The skill follows security best practices by providing an
ask_secrethelper to mask input during the process and stores these values in local.envfiles, which is a standard and recommended practice for local development secrets.
Audit Metadata