skills/boazy/skills/d2-diagrams/Gen Agent Trust Hub

d2-diagrams

Fail

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill includes instructions to install the D2 compiler using a piped shell command: curl -fsSL https://d2lang.com/install.sh | sh. This executes code directly from a remote server without local verification.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes declarative text (D2 syntax) provided by the user or from files, which is then interpreted by the d2 CLI tool. This represents an indirect prompt injection surface where malicious inputs could attempt to influence the agent or exploit the rendering engine.
  • Ingestion points: Processes .d2 files and user-provided syntax snippets.
  • Boundary markers: Uses standard Markdown code blocks to separate syntax instructions.
  • Capability inventory: Utilizes Bash(d2 *) for rendering, along with file read/write permissions (Read, Write, Grep, Glob).
  • Sanitization: No evidence of sanitization or validation of the ingested D2 syntax is present.
  • [EXTERNAL_DOWNLOADS]: The skill fetches content from external domains, including the installation script from d2lang.com, Go packages from oss.terrastruct.com, and a large (~140 MB) Chromium bundle required for PNG exports. It also allows the rendering of icons from arbitrary external URLs.
  • [COMMAND_EXECUTION]: The skill relies on the execution of the d2 command-line utility. While restricted by the allowed-tools configuration, the tool processes complex, user-controlled input and can perform network operations and file writes.
Recommendations
  • HIGH: Downloads and executes remote code from: https://d2lang.com/install.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 23, 2026, 07:07 AM