diagramming
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a bundled Node.js script (
scripts/inspect-excalidraw.mjs) to perform structural validation of Excalidraw scenes, checking for geometry errors and broken element references within diagram files.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes external inputs such as source code and documentation to generate visual diagrams, which establishes a surface for indirect prompt injection. However, the impact is limited to the generation of diagram metadata and structure within the Excalidraw format.\n - Ingestion points: Authoritative code, project documents, and existing
.excalidrawfiles as specified in the skill instructions.\n - Boundary markers: No explicit instructions are provided to the agent to distinguish between or ignore instructions embedded within the source materials being diagrammed.\n
- Capability inventory: The agent is authorized to read local files, write to specified diagram directories (e.g.,
~/code/diagrams), and execute the bundled local validation script.\n - Sanitization: The bundled
scripts/inspect-excalidraw.mjsscript validates the Excalidraw JSON schema and ensures numeric geometry is finite before completion.
Audit Metadata