| name | deep-exploration-01 |
| description | Conduct deep research and multi-perspective exploration for complex questions, decisions, and problem spaces. Use when the user asks for rigorous analysis, deeper context, strategic options, premise checking, thesis testing, creative but defensible alternatives, or a nuanced answer that goes beyond summary and must separate sourced facts from inference. |
Deep Exploration 01
Use this skill to turn broad, ambiguous, or high-complexity prompts into evidence-based analysis with strong premise checking, multiple expert lenses, and explicit verification.
Workflow
- Calibrate the task.
- Identify the real objective, output shape, time sensitivity, and domain boundaries.
- If the prompt contains a material ambiguity, flawed premise, or missing constraint, state it plainly and ask a targeted clarification question.
- If the prompt is workable, proceed without ceremony.
- Build an expert council.
- Choose 3 to 5 expert lenses appropriate to the task.
- Always include one skeptical or adversarial lens that attacks assumptions and weak evidence.
- Keep the lenses domain-relevant; avoid theatrical persona naming.
- Map the solution space.
- Start with the conventional approach.
- Generate three non-obvious alternatives, ideally by blending the problem with a distant but useful domain.
- Evaluate the tradeoffs and choose the best path or hybrid.
- Gather and separate evidence.
- Prefer primary sources and current information when stakes or freshness matter.
- Distinguish clearly between sourced facts, reasoned inference, and speculation.
- Surface what would change the conclusion.
- Run structured challenge.
- Let the skeptical lens challenge the leading approach.
- Let the other lenses respond with counter-evidence or constraints.
- Synthesize the debate into a single working thesis.
- Verify before finalizing.
- Generate fact-checkable verification questions for the main claims.
- Answer them from available evidence.
- Revise the response if verification exposes gaps or overstatement.
Output Contract
Return results in this order when the task is substantial:
Premise Check
Expert Lenses
Candidate Approaches
Chosen Thesis
Evidence and Verification
Final Synthesis
Open Questions
For smaller tasks, compress the same logic into fewer sections.
Guardrails
- Do not present hidden chain-of-thought; present conclusions, evidence, assumptions, and tradeoffs.
- Do not claim certainty where the evidence is thin.
- Do not use "omniscient" or authority-theater language in the final response.
- Do not confuse creativity with rigor; novel options must still be defensible.
- If information is not from the provided sources or verified research, label it as inference or background knowledge.
Research Standard
- Use web research when freshness, citations, or precise source attribution matter.
- Prefer primary documentation, official publications, research papers, and direct data sources.
- Include links when the user would benefit from source access.
Companion Use
- Pair with
timeline-traverser when the user needs explicit future-state simulations across several choices after the evidence is gathered.
- Use this skill for the research and thesis work; use
timeline-traverser for the causal scenario map.
Context Strategy
Load progressively:
- Load this
SKILL.md.
- Load Exploration output patterns for response scaffolds.
- Load Expert lens and verification playbook when the task needs structured debate or deeper rigor.
References