wizard
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates the generation and local execution of bash scripts. The provided template uses standard system utilities (grep, mktemp, tput) and the GitHub CLI (gh) to manage configuration and secrets in a consistent, predictable manner.\n- [PROMPT_INJECTION]: The skill is subject to an indirect prompt injection surface because it reads repository content—such as README files and environment examples—to determine the steps for the generated wizard. Malicious instructions placed in these files could attempt to influence the agent to include deceptive steps.\n
- Ingestion points: The agent is instructed to read .env, .env.example, README, and .github/workflows/* within the target repository.\n
- Boundary markers: There are no explicit instructions for the agent to use delimiters or ignore embedded instructions when processing repository data.\n
- Capability inventory: The generated script possesses capabilities for file system writes (.env), opening browser URLs (open_url), and modifying GitHub repository secrets and variables (gh secret set, gh variable set).\n
- Sanitization: No specific sanitization or filtering logic is prescribed for the data ingested from the repository files during the scoping phase.
Audit Metadata