The Agent Skills Directory

Local File Access for Contextual Awareness: The skill is configured to read from a specific configuration file located at ~/.claude/plugins/config/claude-for-legal/law-student/CLAUDE.md. This is a design feature intended to personalize feedback based on the user's skill level and past patterns. It does not attempt to access sensitive system directories or credentials.
Feedback Persistence and Session Tracking: The skill appends session summaries to a local tracker.md file. This mechanism allows for long-term pattern detection in the student's writing. This is an expected behavior for a pedagogical tool and does not constitute a persistence-based attack vector as it does not modify system startup scripts or executable files.
Input Ingestion Considerations: The skill processes untrusted user data in the form of legal drafts. While this introduces a surface for indirect prompt injection, the risk is effectively mitigated by the skill's limited capabilities; it only generates textual feedback and updates a local log file without performing network requests or shell command execution.

legal-writing