setup-verification
Pass
Audited by Gen Agent Trust Hub on Nov 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze the repository codebase ("Interview the repo") to identify surfaces, run commands, and data structures. This ingestion of untrusted local data creates a potential surface for indirect prompt injection if the repository contains malicious instructions designed to influence the generated driver's logic. 1. Ingestion points: Project repository files analyzed during Step 1 of the skill instructions. 2. Boundary markers: None specified for protecting the agent during repo analysis. 3. Capability inventory: Execution of app driver commands (launch, doctor, cleanup, journey, sql), shell script execution (check-verify-skill.sh), and Playwright journeys. 4. Sanitization: None specified for data extracted from the repository code.
- [COMMAND_EXECUTION]: The provided verification script 'scripts/check-verify-skill.sh' executes the generated driver CLI using the command './app --help'. This is a standard verification step but involves executing a file created during the skill's setup process.
- [DYNAMIC_EXECUTION]: The skill facilitates the generation and runtime execution of local automation code, specifically Playwright journey scripts and a driver CLI ('app'). This is the primary purpose of the skill for providing verifiable proof of application behavior.
Audit Metadata