The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes existing markdown files and other project documentation to populate the dashboard shell.
Ingestion points: The commands/bootstrap-project-console.md file instructs the agent to inventory current project documents including plans, tasks, artifacts, and changelogs.
Boundary markers: Absent. There are no instructions for the agent to use delimiters or to ignore embedded instructions within the ingested data.
Capability inventory: The skill performs file system writes to create and update the project-console.html dashboard file.
Sanitization: No validation or escaping is performed on the ingested data before it is written into the HTML data object.
[DYNAMIC_EXECUTION]: The generated project-console.html file uses unsafe methods to render data derived from external project files.
The script within templates/project-console.html uses .innerHTML to render the vision, plan, todo, artifacts, and changelog sections.
This architecture allows for HTML injection and Cross-Site Scripting (XSS) if the source documentation files contain malicious script tags or event handlers, which would execute in the browser when the dashboard is opened.

html-project-console