markdown-mermaid-writing
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill package is composed entirely of Markdown documentation, style guides, and templates designed to assist an agent in generating technical reports and diagrams. No scripts, binaries, or automated installation processes were detected.
- [SAFE]: External references within the skill are limited to reputable technical documentation (e.g., mermaid.js.org, github.blog) and peer-reviewed scientific literature (e.g., Nature Protocols, Nature Methods). These are presented as standard citations without any dynamic execution or exfiltration patterns.
- [SAFE]: The 'allowed-tools' section in the frontmatter correctly scopes the skill's environment to standard file operations and shell access, and no instructions were found that attempt to exploit these capabilities for unauthorized actions.
- [INDIRECT_PROMPT_INJECTION]: The skill defines an Agent Operating Procedure (AOP) for processing user-provided drafts and data. While this creates a theoretical attack surface for indirect prompt injection, the skill includes explicit 'Integrity rules' that mandate citation verification and user approval for all claims, providing sufficient guardrails for its intended function.
Audit Metadata