The Agent Skills Directory

[SAFE]: The skill manages local application data at ~/.claude/learning/profile.md. This is standard behavior for profile configuration and does not involve accessing sensitive system credentials or unauthorized file paths.
[PROMPT_INJECTION]: The skill identifies a potential surface for indirect prompt injection via the processing of user-pasted feedback in Step 8 of SKILL.md. Ingestion Point: User feedback pasted in the optional final step of profile setup. Boundary markers: None. Capability inventory: File-write to profile.md. Sanitization: The skill implements a manual security control by instructing the agent to extract key themes and present them to the user for confirmation before any data is written to the profile file.

init-learning