memory-clarity-probe
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill provides instructions for the agent to assess text clarity and does not utilize any dangerous tools, network access, or file system operations.
- [PROMPT_INJECTION]: The skill demonstrates a secure pattern for handling untrusted data by using boundary markers to prevent indirect prompt injection.
- Ingestion points: The
memory_contentvariable in Steps 2 and 3 serves as the entry point for external data. - Boundary markers: The skill explicitly uses
<memory>tags to encapsulate untrusted content, isolating it from the agent's instructions. - Capability inventory: No high-risk capabilities (subprocesses, code execution, network requests, or file writes) are available to this skill.
- Sanitization: Content is isolated via XML-style delimiters as a best practice.
Audit Metadata