The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, including textbook chapters, class notes, and case summaries. This creates an indirect prompt injection surface where instructions embedded in these materials could attempt to influence the agent's behavior. However, the skill includes strong internal guardrails, such as 'Hard Rules' against generating content from its own knowledge and the use of Socratic questioning, which mitigate the risk of the agent being manipulated into bypassing its intended purpose.
[SAFE]: The skill reads configuration data from ~/.claude/plugins/config/claude-for-legal/law-student/CLAUDE.md. This file access is scoped to the plugin's own configuration directory and is used to maintain user-specific formatting preferences and track existing outlines. This is a legitimate use of local configuration and does not involve sensitive data exposure or network exfiltration.

outline-builder