skills/ikkeseb/skills/excalidraw/Gen Agent Trust Hub

excalidraw

Pass

Audited by Gen Agent Trust Hub on Sep 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The Node.js utility scripts/excalidraw.mjs executes local browser binaries (such as Chrome, Chromium, or Edge) in headless mode to rasterize SVG layout diagnostics. This execution is limited to local 'file://' URLs, does not involve network access, and is a core functional requirement of the preview workflow.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process externally provided diagram specifications in JSON format. It mitigates potential injection and schema confusion risks through a multi-stage validation pipeline in compileScene and validateScene that enforces strict typing, allowed keys, and reciprocal element bindings.
  • Ingestion points: The skill reads JSON files from the local file system based on user-supplied paths.
  • Boundary markers: The skill documentation mandates distinct validation and inspection states before delivery.
  • Capability inventory: File system read/write operations and local command execution for browser-based rendering.
  • Sanitization: Input data is validated against a closed vocabulary and strict numeric constraints.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 24, 2026, 09:27 AM