The Agent Skills Directory

[DATA_EXPOSURE]: The skill constructs file system paths using a user-provided course-slug argument without explicit validation markers in the instruction set.
Evidence: The paths docs/instructional-design/courses/{slug}/course.json and docs/instructional-design/courses/{slug}/course.html rely on the {slug} variable.
Risk: An unsanitized input could potentially be used for path traversal, although the impact is limited by the skill's specific directory scope.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from a JSON file to generate an HTML visualization, creating a vulnerability surface for indirect injection.
Ingestion points: The file docs/instructional-design/courses/{slug}/course.json (SKILL.md).
Boundary markers: None present in the instructions to delimit the data for the sub-agent.
Capability inventory: The skill utilizes an internal agent to read local files and write generated HTML files back to the filesystem.
Sanitization: No sanitization or validation of the JSON content is mentioned before it is passed to the course-visualizer.md agent or rendered into HTML.

course-visualize