creating-diagrams

Pass

Audited by Gen Agent Trust Hub on Oct 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads the D2 diagramming tool from its official repository on GitHub. The download is performed via a dedicated setup command and includes verification against hardcoded SHA-256 checksums to ensure file integrity.
  • [COMMAND_EXECUTION]: The skill executes external commands to manage its environment and render diagrams. It uses the system's tar or unzip utilities to extract the downloaded D2 binary and subsequently runs the binary itself to process diagram source files. These executions use fixed arguments and do not utilize shell-sensitive features like command chaining or redirection.
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts D2 source code as input, which represents an attack surface for indirect instructions.
  • Ingestion points: D2 source files read in scripts/creating-diagrams.mjs.
  • Boundary markers: Not used.
  • Capability inventory: Spawns the D2 binary for rendering; writes output files to the local filesystem.
  • Sanitization: The script performs static analysis on the D2 source code to detect and reject direct styling overrides, ensuring that diagrams conform to a secure and accessible set of predefined roles.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 15, 2026, 07:18 AM