madman
Fail
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: CRITICALNO_CODEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill contains no executable code, scripts, or automated tool invocations. It operates solely as a set of behavioral instructions for an AI agent.\n- [NO_CODE]: There are no scripts or binaries included in this skill; all functionality is defined through natural language instructions.\n- [EXTERNAL_DOWNLOADS]: The skill contains references to documentation and other skills on the author's GitHub repository. These are informational links and do not involve automated or unrequested downloads. The URLs originate from the author's verified vendor infrastructure.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied 'seeds' (topics, experiences, and intent) to generate writing material. This ingestion point is low-risk because the skill lacks capabilities to perform sensitive operations like file system modification or network exfiltration.\n
- Ingestion points: User-provided 'seed' containing topics and lived experiences (SKILL.md, Step 1).\n
- Boundary markers: The output template requires the AI to summarize the seed at the start of the output, creating a structural separation between input and generated content.\n
- Capability inventory: None. No subprocess calls, file writes, or network operations are present in any of the skill's files.\n
- Sanitization: Instructions focus the AI on the user's unique experience and editorial intent without requiring specific technical sanitization code.
Recommendations
- Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata