marp-deck-visual-designer
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill uses localized shell commands to manage its environment and asset pipeline. This includes checking for configuration files (
ls package.json), managing dependencies (npm install), and executing rendering workflows via a task runner (task mermaid,task excalidraw-file). These operations are explicitly tied to the skill's primary purpose of generating presentation assets. - [EXTERNAL_DOWNLOADS]: Fetches the
@napi-rs/canvaspackage from the official NPM registry. This is a well-known library used for server-side image rendering and is required for converting Excalidraw JSON into PNG images. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-controlled data from
planning/deck-plan.mdto generate diagrams and layout instructions. While this represents an attack surface, the risk is mitigated by the fact that the output is rendered into static images or Markdown. - Ingestion points: Reads content from
planning/deck-plan.mdto derive visual structures. - Boundary markers: No explicit delimiters are used to separate user data from system instructions during the materialization phase.
- Capability inventory: Includes file system writes, NPM package installation, and execution of local scripts and task commands.
- Sanitization: No specific sanitization or validation of the input Markdown is described; the skill assumes the plan adheres to the expected structural format.
Audit Metadata