create-mermaid-diagrams
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The workflow in
SKILL.mdinstructs the agent to use the Mermaid CLI (mmdc) for diagram validation via the command:mmdc -i diagram.mmd -o /tmp/diagram-check.svg. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted user descriptions to generate Mermaid source code, which creates a potential surface for indirect prompt injection.
- Ingestion points: User-provided relationships, sequences, and state descriptions processed during the workflow defined in
SKILL.md. - Boundary markers: None explicitly defined in the instructions for delimiting user input.
- Capability inventory: Subprocess execution via the
mmdcCLI tool as described inSKILL.md. - Sanitization: The skill includes a basic formatting requirement to "Quote labels containing punctuation," which provides minimal structural sanitization for the Mermaid renderer.
Audit Metadata