game-testing
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The startServer helper function provided in the Quick Start section of SKILL.md is vulnerable to path traversal.
- Evidence: The function uses path.join(dir, req.url) without sanitizing the input URL, which could allow unauthorized access to local files outside the intended directory if the test server is accessed maliciously.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting and processing untrusted data from the browser environment.
- Ingestion points: The skill captures data from the game being tested via page.evaluate(), page.on('console'), and page.on('pageerror') calls found in SKILL.md.
- Boundary markers: No delimiters or instructions are used to prevent the agent from interpreting content from the game state or logs as commands.
- Capability inventory: The skill includes file system access (fs) and browser automation capabilities (playwright).
- Sanitization: No input validation or output escaping is performed on the data retrieved from the browser context before it is logged or processed.
Audit Metadata