mermaid-diagram-pro
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external technical data (codebase logic, database schemas, and workflows) to generate diagrams, creating a surface for indirect instructions to influence agent behavior.
- Ingestion points: The skill gathers technical details, symbols, state transitions, and relationship cardinality from the codebase (SKILL.md, Protocol Step 2).
- Boundary markers: There are no explicit instructions or delimiters provided to separate ingested codebase content from the agent's internal logic or instructions.
- Capability inventory: The skill references local execution of linting and generation scripts (
scripts/verify-mermaid.shandscripts/schema-to-erd.ts) that interact with the provided codebase data. - Sanitization: The instructions lack explicit sanitization or escaping of external content before interpolation into Mermaid syntax. Furthermore, the skill suggests configuring the Mermaid
securityLeveltolooseinreferences/styling-themes.md, which enables HTML rendering in labels and potentially increases the impact of malicious content embedded in the ingested codebase.
Audit Metadata