plain-visual-docs
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to use
npxandnpmto fetch and execute@mermaid-js/mermaid-cliandpuppeteer. These are standard packages from the public npm registry used for rendering diagrams. - [COMMAND_EXECUTION]: The skill uses shell commands to manage a local workspace in the
.scratchpaddirectory and execute the diagram rendering pipeline. - [DYNAMIC_EXECUTION]: The skill employs a Python heredoc (`python3
- <<'EOF'`) to dynamically parse Markdown files and extract Mermaid code blocks for validation.
- [INDIRECT_PROMPT_INJECTION]: The validation logic processes content from external Markdown files, which represents an ingestion surface for potential instructions embedded in data. • Ingestion points: Reads file content via
open("PATH/TO/DOC.md").read()in a Python script. • Boundary markers: Absent; parsing relies on standard Markdown code fences. • Capability inventory: Executes shell commands (npx,npm) and writes files to the local file system. • Sanitization: Absent; content inside diagrams is passed directly to the rendering tool.
Audit Metadata