page-cro
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions define an 'ExecutionProxy Interface' (using the
rtkprefix) for performing terminal operations. Specifically, it mentions executing commands likertk npm testas part of a Test-Driven Development (TDD) protocol. While this involves command execution, it is framed within a legitimate software development and verification workflow. - [INDIRECT_PROMPT_INJECTION]: The skill explicitly acknowledges the potential for indirect prompt injection by stating that web content and external data (processed via
BrowserOS) should be treated as 'hostile.' It provides high-level guidance to redact sensitive information (PII/secrets) before sharing context with subagents, addressing the inherent risk surface of processing untrusted web content.
Audit Metadata