The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from repository markdown files and user context to generate HTML visualizations.
Ingestion points: Markdown files, docs/SUMMARY.md, phase-XX-*.md, and user-provided context as defined in SKILL.md and references/workflow.md.
Boundary markers: Uses VISUALIZE: placeholders in templates (e.g., references/templates/context.html) to delimit where source content is placed.
Capability inventory: Reading repository files and writing HTML/CSS files to the local file system (SKILL.md).
Sanitization: Absent; the instructions do not specify a requirement to sanitize or escape HTML/JS content from the source files before placing it into the final visualization.
[EXTERNAL_DOWNLOADS]: The skill's HTML templates in references/templates/ include a script module that fetches the Mermaid library from the official jsDelivr CDN.
Evidence: https://cdn.jsdelivr.net/npm/mermaid@11/dist/mermaid.esm.min.mjs is referenced in context.html, document.html, and plan.html.
Source: Fetched from a well-known content delivery network.

visualize