The Agent Skills Directory

[SAFE]: The skill implements a local persistent memory system by writing markdown files to a project-specific directory (/home/roger/.claude/projects/). This functionality is restricted to the user's local environment and aligns with the stated purpose of maintaining architectural and behavioral context across sessions.\n- [SAFE]: A multi-step verification process is enforced, requiring the agent to abstract conversation patterns into clear rules and obtain explicit user confirmation before finalizing any file writes. This human-in-the-loop requirement effectively mitigates the risk of persisting incorrect or malicious instructions.\n- [SAFE]: The skill establishes clear data boundaries by explicitly prohibiting the storage of ephemeral task details, code snippets, or sensitive personal judgments, preventing the accumulation of redundant or inappropriate information.

continuous-learning