The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external entities such as personas, competitors, and customer proof points, then interpolates this content directly into HTML templates for interactive lesson slides.
Ingestion points: Data is retrieved from the workspace environment via tools including list_entities, get_entity, and find_motion_icp as specified in the task method.
Boundary markers: There are no explicit instructions for the agent to use boundary markers or to ignore embedded instructions when processing these external data points.
Capability inventory: The skill has the capability to write local HTML files and include JavaScript functionality (assets/lesson.js) that is executed in the user's browser.
Sanitization: The instructions do not explicitly require the agent to sanitize or escape the retrieved data before rendering it into the final HTML output, creating a potential surface for injection.
[EXTERNAL_DOWNLOADS]: The generated HTML output references Google Fonts as an external dependency for visual styling.
This reference targets a well-known service and is used solely for typography purposes within the generated primer.
[DYNAMIC_EXECUTION]: The skill generates interactive artifacts by inlining local JavaScript and CSS templates into self-contained HTML files at runtime.
The agent is instructed to assemble the assets/lesson.js script and CSS into a single portable file. This represents a standard templating task using pre-defined local assets.

train