cocounsel-legal:deep-research

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFE
Full Analysis
  • External Service Integration: The skill interacts with the cocounsel-legal MCP server to access Westlaw's legal databases. This facilitates the retrieval of legal authorities and the synthesis of research reports from established legal sources.
  • Polling and Process Management: To manage long-running research tasks, the skill uses standard polling logic with backoff intervals implemented via the Bash sleep command. This ensures the agent waits appropriately for the research to complete without unnecessary service overhead.
  • Information Ingestion Surface: The skill is designed to retrieve and present legal reports from an external provider. Ingestion points: Untrusted data enters the agent context through the legal_research_get_deep_research_report tool output (Step 4). Boundary markers: The skill does not currently use specific delimiters or boundary markers for this external content, instead prioritizing data integrity by instructing the agent to provide the report verbatim. Capability inventory: The agent's available capabilities include the Bash sleep command and specialized MCP tools for the legal research workflow. Sanitization: The content is presented exactly as received from the research provider. This configuration is consistent with the primary purpose of delivering accurate, unedited legal research information from a recognized professional service.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 12:28 AM
Security Audit — agent-trust-hub — cocounsel-legal:deep-research