worker-classification

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill does not contain any malicious patterns such as prompt injection, obfuscation, or unauthorized data exfiltration. All identified behaviors are consistent with its stated purpose as a legal worker classification tool.
  • [DATA_EXPOSURE]: The skill accesses local configuration and matter-specific files within a dedicated plugin directory (~/.claude/plugins/config/claude-for-legal/). This behavior is intended for maintaining jurisdictional footprints, escalation tables, and matter workspace context.
  • [COMMAND_EXECUTION]: The skill interacts with external legal research tools (such as Westlaw, CourtListener, or configured MCP connectors) to fetch currently operative classification tests. This is a primary function of the skill and uses neutral, well-known service integrations.
  • [PROMPT_INJECTION]: The skill includes strong instructional guardrails, such as a mandatory 'prospective-only hard gate' that prevents out-of-scope usage for existing arrangements unless explicitly overridden by the user with a mandatory warning banner.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 03:58 PM
Security Audit — agent-trust-hub — worker-classification