The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data (course materials and verified manifests) and user-controlled files to generate homework suggestions, creating an injection surface.
Ingestion points: SKILL.md describes workflows that read from a "verified manifest" (Step 1), summary_md (Step 2), and user-supplied files such as profile.md, goals.md, and projects.md (Step 3).
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" wrappers to isolate content processed from external manifests or user files.
Capability inventory: The skill is authorized to read and write files within the ~/goal-driven-ai-context/ directory, including creating new homework records and updating progress logs.
Sanitization: No explicit sanitization or validation steps are defined for the incoming text data before it is interpolated into the agent's reasoning process.
[PERSISTENCE]: The skill implements functional persistence to maintain state across coaching sessions.
The agent is instructed to create and update files in a specific directory (homework/*.md and progress.md) to track user commitments and achievements. This is consistent with the skill's primary purpose and does not indicate unauthorized system persistence.

lab-homework