The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill accepts natural language input from the user to describe course changes (Paso 2).
Ingestion points: User-provided change descriptions in SKILL.md Step 2.
Boundary markers: Not explicitly defined in this instruction set.
Capability inventory: File write operations (course.json, CHANGELOG.md) and internal agent dispatching (changelog-generator.md, cmi5-metadata-writer.md).
Sanitization: No explicit sanitization or validation of the text content is mentioned within these instructions, though the workflow includes validation agents in Steps 4 and 5.
[COMMAND_EXECUTION]: The skill performs file system writes to update course configuration and maintenance logs (Paso 7). These operations are restricted to the local course directory defined by the slug parameter and represent the primary functionality of the tool.

course-revise