markdown-mermaid-writing
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines best practices for Markdown and Mermaid authoring, focusing on accessibility, diagram syntax validation, and visual verification of artifacts.\n- [EXTERNAL_DOWNLOADS]: The skill utilizes vendor-owned tools (sandbase_discover, sandbase_inspect, sandbase_run) to access the SandBase API for media generation. The workflow ensures transparency by requiring the agent to display pricing and obtain explicit user confirmation before executing external tasks.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data by inspecting user files and source materials (SKILL.md). While it lacks explicit boundary markers or input sanitization, the risk is mitigated by the requirement for manual output verification and user approval for subsequent tool actions (sandbase_run).
Audit Metadata