draw-io

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes a shell script scripts/convert-drawio-to-png.sh that executes the drawio command to export diagrams to PNG format and the git add command to stage the resulting images. These actions are aligned with the skill's primary purpose of automating diagram workflows and occur within the project context.
  • [INDIRECT_PROMPT_INJECTION]: The skill operates on .drawio XML files provided by the user or external sources, creating a potential surface for indirect prompt injection if the files contain adversarial content. However, the risk is mitigated by the skill's focused application on diagram properties and the use of the drawio utility for processing.
  • Ingestion points: The agent reads and modifies .drawio XML files during the layout adjustment and conversion process (SKILL.md, README.md).
  • Boundary markers: No explicit delimiters or boundary markers are defined to isolate diagram data from potential instructions embedded within the XML tags.
  • Capability inventory: The skill utilizes drawio and git commands through the convert-drawio-to-png.sh script to process and stage file changes.
  • Sanitization: Content within the XML files is not explicitly sanitized by the provided scripts, relying on the agent's logic and the underlying drawing tools for safe handling.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 12:59 PM