ac-diagram
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (source code files and business documentation) to generate visual models. This creates a surface for indirect prompt injection if the ingested files contain malicious instructions designed to influence the agent's output. However, the skill explicitly defines a 'read-only' constraint for the analysis phase and implements manual checkpoints for artifact generation.
- Ingestion points: The skill reads modules, files, functions, and business process descriptions via the
ace-toolandExploresubagents (SKILL.md, analyzer.md). - Boundary markers: Absent for ingested content.
- Capability inventory: File writing and URL generation via
mermaid-liveMCP; repository traversal viaExploretool. - Sanitization: Not explicitly defined in the instructions.
- [EXTERNAL_DOWNLOADS]: The skill uses the
mermaid-liveMCP to export diagrams or generate visualization URLs. This is a functional requirement for the skill's stated purpose of diagram generation. - [COMMAND_EXECUTION]: The skill facilitates the export of local files (SVG, PNG, PDF) using the
mermaid-liveMCP. This capability is scoped to the output of generated diagrams and requires user-specified format parameters.
Audit Metadata