draw-io
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions and local scripts for converting diagrams to images and searching for icons.
- The shell script
scripts/convert-drawio-to-png.shuses thedrawioCLI to export XML to PNG and automatically stages the resulting images in the local Git repository usinggit add. - The Python script
scripts/find_aws_icon.pyperforms local searches within the skill's reference files. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and modify
.drawioXML files provided by the user. - Ingestion points: The agent reads and parses external XML diagram files.
- Capability inventory: The skill can execute local scripts, perform file system writes (generating PNGs), and interact with Git commands.
- Sanitization: Instructions focus on precise manipulation of specific XML tags (
mxGeometry,mxCell) and coordinate calculations, which limits the risk of arbitrary command injection through the diagram data. - [EXTERNAL_DOWNLOADS]: The documentation references external resources and repositories on GitHub (e.g.,
jgraph/drawio,m-radzikowski/diagrams-aws-icons) for icon assets and general documentation. These references are informative and point to well-known community resources.
Audit Metadata