The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data by scanning an external "output folder" containing documents such as Privacy Impact Assessments (PIAs), Data Processing Agreements (DPAs), and Subject Rights Request (SRR) replies. These documents could contain malicious instructions designed to influence the agent's gap analysis or policy language generation.
Ingestion points: The skill reads arbitrary files from a directory defined in ~/.claude/plugins/config/claude-for-legal-zh/privacy-legal/CLAUDE.md.
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded content within the processed documents.
Capability inventory: The skill has permissions to read local files and write updates back to its own configuration file.
Sanitization: No explicit sanitization or filtering is applied to the legal text extracted from documents before it is used to generate suggestions.
[DATA_EXPOSURE]: The skill specifically targets Subject Rights Request (SRR) replies and Data Processing Agreements (DPAs) for analysis. While this is the intended functionality, these documents typically contain sensitive personal identifiable information (PII) and confidential contractual terms. The skill maintains a history of scan dates and paths within its configuration directory.

policy-monitor