comprehensive-research-agent

To prevent silent failures during multi-source research, enforce explicit validation checkpoints and error recovery protocols. This approach mandates handling tool failures with fallback strategies and clear gap documentation to ensure transparent reasoning and traceability.

Cross-source validation is the process of cross-checking key claims against multiple fetched sources to prevent hallucinations. It explicitly notes agreements, conflicts, and missing information while maintaining a catalog of fetched versus referenced sources for audit trails.

Structure multi-step information gathering by using substantive thinking blocks that document insights, connections, gaps, and decision rationales at each phase transition. Apply validation checkpoints to confirm tool outputs, source relevance, and information completeness before proceeding.

No specific tool dependencies are required for this research approach. It applies to complex information-gathering tasks that involve tool calls, cross-source validation, and decision logging, functioning effectively without external dependencies.

The best way to ensure research completeness is to apply guardrails with a pre-completion checklist. This verifies coverage of required topics and sources, ensuring explicit validation steps are met before declaring the multi-source research synthesis complete.

You should avoid structured thinking blocks for research synthesis when a task requires simple, single-source retrieval without complex decision logging. This approach is designed for multi-step information gathering needing explicit error handling and cross-source validation.