The Agent Skills Directory

Local File System Interaction: The skill reads from and writes to specific paths within the ~/.claude/plugins/config/claude-for-legal/ directory. This is used to manage matter-specific context and store work products, which is consistent with its stated purpose of summarizing legal diligence. No access to sensitive system files or credentials outside this directory was observed.
Indirect Data Processing: This skill processes outputs from the diligence-issue-extraction skill. While this creates a dependency on external data, the risk is minimal as the skill only performs summarization and formatting tasks without executing code or reaching out to the network.
No External Connectivity: The skill does not contain any commands for network operations such as fetching remote scripts or exfiltrating data. It operates strictly on the local data provided in its configured workspace.
Defined Operational Boundaries: The instructions include clear constraints on what the skill should and should not do, such as not making autonomous materiality calls and requiring human review before distribution, which aligns with security best practices for AI agents.

deal-team-summary