e2e-reviewer
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted test files provided by the user, such as E2E specifications and Page Object Models. Malicious instructions hidden within these files could potentially influence the agent's behavior during the Phase 2 LLM analysis step.\n
- Ingestion points: User-provided E2E test files and Page Object Model files (e.g., those located in the
e2e/directory).\n - Boundary markers: Absent. The skill instructions do not specify the use of delimiters or 'ignore' warnings to isolate external data from the analysis instructions.\n
- Capability inventory: The skill primarily uses a
Greptool for read-only operations and analysis via the LLM; no dangerous capabilities like file system writes or network operations are utilized by this skill.\n - Sanitization: Absent. The user-provided code is reviewed directly by the LLM without validation or escaping.
Audit Metadata