The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its processing of external and session-based data to define course objectives and justifications.
Ingestion points: The agent reads content from transcript.md, .audit/ files, session records, .mentor/ records, learner.md, and findings from the /research subagent.
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat ingested data as distinct from instructions or to disregard embedded directives.
Capability inventory: The agent has the authority to write unit.md files (defining outcomes for the teacher and auditor), update roadmap.md, generate roadmap.html, and trigger additional research subagents.
Sanitization: The skill does not specify any validation, filtering, or escaping of the ingested content before it is used to generate the next planning documents.

advisor