docs-maker
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements robust defensive instructions to prevent prompt injection and data poisoning. It explicitly instructs the agent to treat retrieved content as evidence rather than authoritative instructions, and to disregard commands like 'ignore previous instructions' embedded in external data.
- [EXTERNAL_DOWNLOADS]: The skill references official documentation and developer platforms for well-known services, specifically Anthropic (docs.anthropic.com, platform.claude.com) and OpenAI (developers.openai.com, platform.openai.com). These references are used for technical guidance and do not represent a security risk.
- [PROMPT_INJECTION]: Deterministic flags regarding instruction overrides are false positives. The skill defines internal authority hierarchies (User/Project > Retrieved Material) and includes guidance for evaluating agent behavior against adversarial inputs, which is a defensive capability.
- [DATA_EXFILTRATION]: No sensitive file access or network-based data exfiltration patterns were detected. The skill operates purely as a text and document refactoring tool.
Audit Metadata