reg-gap-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Local Configuration Access: The skill is designed to read from
~/.claude/plugins/config/claude-for-legal/privacy-legal/CLAUDE.mdto establish the current state of a company's privacy commitments and DSAR systems. While this involves accessing sensitive internal data, it is the core mechanism for performing the intended gap analysis. - Indirect Prompt Injection Surface: The skill ingests regulatory text or summaries provided by users or retrieved from external research tools. This creates a surface for indirect prompt injection, where malicious instructions embedded in the regulatory text could attempt to manipulate the agent's gap analysis findings or remediation plan.
- Ingestion points: Regulatory text or summaries provided as arguments or via research tools.
- Boundary markers: The instructions lack explicit delimiters to separate the untrusted regulatory text from the agent's internal logic, though it does categorize the requirements into a table.
- Capability inventory: The agent can read local files (
CLAUDE.md) and generate new documentation based on the analysis. - Sanitization: There is no mention of sanitizing or escaping the content of the regulation before processing.
- External Research Connectors: The skill instructions reference external legal research tools (e.g., Westlaw, EUR-Lex). These are utilized to verify citations and ensure the currency of regulatory data, which is standard practice for legal professionals using AI assistants.
Audit Metadata