deep-research
This skill should be used when the user needs deep, multi-step research using native web tools (WebSearch/WebFetch) without external API keys.
Deep Research (No External API)
Purpose
Run structured, multi-step research with evidence-first synthesis.
This skill is designed to work without Google/Gemini API or any paid provider setup.
When to Use This Skill
Use this skill when:
- Performing market analysis
- Conducting competitive landscaping
- Building literature or source reviews
- Doing technical due diligence
- Preparing decision memos with citations
Requirements
- Access to built-in web research tools (
WebSearch,WebFetch) - Clear research question and scope
No external API key is required.
Progress Tracking
Display a progress gauge at each research phase:
[████░░░░░░░░░░░░░░░░] 20% — Phase 1/5: Defining Objective & Scope
[████████░░░░░░░░░░░░] 40% — Phase 2/5: Collecting Sources
[████████████░░░░░░░░] 60% — Phase 3/5: Deep Analysis & Synthesis
[████████████████░░░░] 80% — Phase 4/5: Drafting Output
[████████████████████] 100% — Phase 5/5: Quality Review & Citations
Workflow
Research Protocol
- Define objective and output format
- Clarify the research question, audience, and success criteria.
- Confirm time horizon (e.g., last 12 months, last 5 years).
- Build search strategy
- Create 5-10 query variants (broad, narrow, comparative, opposing views).
- Prefer primary sources first (official docs, papers, regulatory pages, vendor docs).
Parallel Query Execution
Do NOT run searches sequentially. Launch one ResearchScout agent per query variant simultaneously in a single block.
| Agent | Query Type |
|---|---|
ResearchScout-Broad | Wide-scope 2–3 keyword query |
ResearchScout-Narrow | Specific 5+ keyword query |
ResearchScout-Comparative | "X vs Y" or "comparison of" framing |
ResearchScout-Opposing | Contrarian or critical viewpoints |
ResearchScout-Primary | Official docs, papers, government sources |
ResearchScout-Recent | Date-filtered to last 12 months |
Each agent prompt begins with:
# ResearchScout — Targeted Web Research Agent
Role: Execute a single search query using WebSearch/WebFetch, collect top results with URLs and key findings, return structured results.
Wait for all ResearchScout agents to complete. Deduplicate results by URL. Then proceed to synthesis/triangulation phase.
- Collect sources with traceability
- Capture URL, title, date, publisher, and relevance notes.
- Track contradictions and unresolved claims explicitly.
- Validate and triangulate
- Cross-check key claims with at least 2 independent credible sources.
- Flag weak evidence and avoid overconfident conclusions.
- Synthesize output
- Produce concise findings with direct citations.
- Separate facts, inferences, and recommendations.
Output Formats
Choose one based on request:
1) Executive Brief
- Objective
- Top findings (5-10)
- Risks / unknowns
- Recommendations
- Sources
2) Comparative Analysis
- Criteria matrix
- Option-by-option strengths/weaknesses
- Trade-offs
- Recommendation + rationale
- Sources
3) Research Log
- Queries used
- Source inventory
- Evidence quality notes
- Open questions
- Next research steps
Quality Bar
- Evidence before conclusions
- Date-aware and source-aware claims
- Contradictions surfaced, not hidden
- No uncited critical claims
Time & Cost
- Time: usually 5-20 minutes depending on scope
- Cost: no external API cost for this skill
Safety
- Never fabricate sources or citations.
- If evidence is insufficient, state it clearly.
- Distinguish confirmed facts from inference.
Critical Rules
- Always include citations for material claims.
- Always separate facts from interpretations and recommendations.
- Always mark uncertainty explicitly when evidence is weak or conflicting.
Error Handling
| Error | Likely Cause | Action |
|---|---|---|
| WebSearch returns no results | Query too specific, misspelled, or topic very niche | Broaden query, try alternate phrasing, report low-coverage finding |
| WebFetch times out or blocked | Site is down, bot-blocking, or paywalled | Skip that source, note it as inaccessible, continue with other sources |
| Insufficient sources found | Topic has limited public information | Report coverage gaps; recommend user provide domain-specific sources |
| Conflicting information across sources | Different sources cite different facts | Flag the conflict explicitly; present both sides with sources |
| Query too broad | Research question covers too many sub-topics | Ask user to narrow the scope or prioritize specific dimensions |
| Paywalled content | Article requires subscription | Note the source as paywalled; use abstract/preview if available |
Example Usage
- "Use deep-research to compare 3 vector databases for enterprise use."
- "Use deep-research to summarize regulatory updates from the last 12 months."
- "Use deep-research to produce a source-backed buy-vs-build memo."