The Agent Skills Directory

[DATA_EXPOSURE]: The skill reads and writes to files located in ~/.claude/plugins/config/claude-for-legal/legal-clinic/. While these files contain sensitive legal case information, all data operations are local to the agent's environment, and there are no network tools or remote URLs utilized for exfiltration.
[COMMAND_EXECUTION]: No shell command execution or system-level operations are present. The skill logic is confined to text processing and structured data management.
[REMOTE_CODE_EXECUTION]: The skill does not download or execute any external scripts or packages.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from external files and other skills (e.g., /client-intake, /draft).
Ingestion points: Reads from ~/.claude/plugins/config/claude-for-legal/legal-clinic/deadlines.yaml and references/plausibility-bands/{省份}.md.
Boundary markers: The skill explicitly requires human confirmation before performing any write operations to the configuration files.
Capability inventory: Limited to file read/write operations within a specific directory.
Sanitization: All data is treated as structured YAML or Markdown for reporting and validation against local 'plausibility bands'.

deadlines