Deep Research

**Judgement:** A strong, well-structured deep research skill that emphasizes methodology, source evaluation, and transparent documentation. Best suited for non-trivial questions where decisions depend on the quality and traceability of evidence. **Strengths** - Enforces a full research cycle (from scoping to delivery) instead of one-shot answers. - Systematically evaluates sources (authority, recency, evidence, bias, corroboration) and tracks them, reducing ungrounded claims. - Provides confidence levels, caveats, and explicit gaps, which is valuable for high-stakes or professional use. - Adjustable depth tiers allow tailoring effort to the problem’s importance. **Key Risks / Limitations** - Can be overkill for simple factual queries, increasing latency and cost versus faster skills. - Quality still depends on the underlying search tools and data; well-documented but still possibly wrong conclusions if sources are weak or biased. - Not optimized for real-time monitoring or rapidly changing events; the protocol is geared toward one-off deep investigations. **Recommended Scenarios** - Strategic, product, or policy decisions where a documented research trail and confidence assessment are required. - Technical or scientific explorations where primary sources and citation trails matter. - Market, competitor, or landscape analyses where conflicting viewpoints and bias need to be surfaced and weighed. - Any situation where the user explicitly wants “deep research” rather than a quick synthesized overview.