The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs file read and write operations constrained to the ~/.claude/plugins/config/claude-for-legal/ directory. These operations are used to retrieve intake data (intake.md), load style guides (CLAUDE.md), and output drafted content (.docx and .md). This is a standard and expected use of file system capabilities for a document generation skill.
[SAFE]: No malicious patterns were identified. The skill does not perform network operations, access sensitive system files (e.g., SSH keys or credentials), or use obfuscation. The instructional logic provides explicit safety gates (confidentiality filtering and risk checks) intended to protect the user's legal position.

demand-draft