Deep Research
Run a citation-backed research pipeline when you need defensible answers on markets, tech, or “X vs Y” before committing to a build.
Overview
Deep Research is an agent skill most often used in Idea (also Validate, Launch) that runs a multi-phase, citation-tracked research pipeline with evidence persistence and structured report output.
Install
npx skills add https://github.com/199-biotechnologies/claude-deep-research-skill --skill deep-researchWhat is this skill?
- Three depth modes: quick (3 phases, ~2–5 min), standard (6 phases, ~5–10 min, default), deep (8 phases, ~10–20 min)
- Citation-tracked pipeline with evidence persistence, claim verification, and progressive context management
- Explicit decision tree routes simple lookups to WebSearch and keeps complex analysis in the research workflow
- Autonomy-first operation with high-materiality assumptions surfaced in Introduction and Methodology
- Structured report generation for comparisons, trends, and state-of-the-art questions
- 3 research depth modes (quick, standard, deep)
- Standard mode: 6 phases (~5–10 min)
- Deep mode: 8 phases (~10–20 min)
Adoption & trust: 7.9k installs on skills.sh; 751 GitHub stars; 2/3 security scanners passed (skills.sh audits).
What problem does it solve?
You need a decision-grade answer backed by multiple sources, but ad-hoc searching loses citations, contradicts itself, and never becomes a report you can reuse.
Who is it for?
Solo builders comparing options, mapping a market or tech landscape, or writing a cited brief before prototyping or positioning.
Skip if: Single-fact lookups, stack traces, or anything the skill itself flags as answerable with one or two searches—use normal web search instead.
When should I use this skill?
User needs multi-source research with citation tracking and structured reports—triggers include “deep research,” “comprehensive analysis,” “research report,” “compare X vs Y,” “analyze trends,” or “state of the art”; not
What do I get? / Deliverables
You get a structured research report with tracked evidence and verified claims at the depth you chose (quick, standard, or deep), ready to inform scope, positioning, or go/no-go calls.
- Structured research report with introduction and methodology
- Citation-tracked evidence and verified claims
- Mode-appropriate depth output (quick, standard, or deep)
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Idea/research is the canonical shelf because the skill’s default job is exploratory evidence gathering with persisted claims—not shipping code or ops. The research subphase fits multi-source synthesis, source identity, and structured reports rather than one-shot search or debugging.
Where it fits
Map competitors and pricing models with cited sources before you pick a niche.
Produce a structured X vs Y report with claim verification instead of opinionated chat.
Compare build approaches (monolith vs services) using standard or deep mode before locking a prototype plan.
Synthesize distribution and AI-search visibility trends with a methodology section you can quote on a landing page.
How it compares
Use instead of chaining raw WebSearch turns when you need persisted evidence, claim verification, and a formal report—not a chat summary.
Common Questions / FAQ
Who is deep-research for?
Solo and indie builders using Claude Code, Cursor, or similar agents who need multi-source, citation-backed analysis for product and strategy decisions—not casual Q&A.
When should I use deep-research?
In Idea/research for landscape and competitor evidence; in Validate/scope when comparing build or pricing paths; at Launch/geo or distribution research when you need cited trends—any time triggers like “deep research,” “compare X vs Y,” or “state of the art” appear.
Is deep-research safe to install?
Review the Security Audits panel on this Prism page for install risk, permissions, and any published audit results before running it on sensitive repos or credentials.
SKILL.md
READMESKILL.md - Deep Research
# Python __pycache__/ *.py[cod] *$py.class *.so .Python # Virtual environments venv/ ENV/ env/ # IDE .vscode/ .idea/ *.swp *.swo *~ # OS .DS_Store Thumbs.db # Research output (kept local) *.json !schemas/*.json # Test output .pytest_cache/ .coverage htmlcov/ --- name: deep-research description: Use when the user needs multi-source research with citation tracking, evidence persistence, and structured report generation. Triggers on "deep research", "comprehensive analysis", "research report", "compare X vs Y", "analyze trends", or "state of the art". Not for simple lookups, debugging, or questions answerable with 1-2 searches. --- # Deep Research ## Core Purpose Deliver citation-tracked research reports through a structured pipeline with evidence persistence, source identity management, claim-level verification, and progressive context management. **Autonomy Principle:** Operate independently. Infer assumptions from context. Only stop for critical errors or incomprehensible queries. Surface high-materiality assumptions explicitly in the Introduction and Methodology rather than silently defaulting. --- ## Decision Tree ``` Request Analysis +-- Simple lookup? --> STOP: Use WebSearch +-- Debugging? --> STOP: Use standard tools +-- Complex analysis needed? --> CONTINUE Mode Selection +-- Initial exploration --> quick (3 phases, 2-5 min) +-- Standard research --> standard (6 phases, 5-10 min) [DEFAULT] +-- Critical decision --> deep (8 phases, 10-20 min) +-- Comprehensive review --> ultradeep (8+ phases, 20-45 min) ``` **Default assumptions:** Technical query = technical audience. Comparison = balanced perspective. Trend = recent 1-2 years. --- ## Workflow Overview | Phase | Name | Quick | Std | Deep | Ultra | |-------|------|-------|-----|------|-------| | 1 | SCOPE | Y | Y | Y | Y | | 2 | PLAN | - | Y | Y | Y | | 3 | RETRIEVE | Y | Y | Y | Y | | 4 | TRIANGULATE | - | Y | Y | Y | | 4.5 | OUTLINE REFINEMENT | - | Y | Y | Y | | 5 | SYNTHESIZE | - | Y | Y | Y | | 6 | CRITIQUE | - | - | Y | Y | | 7 | REFINE | - | - | Y | Y | | 8 | PACKAGE | Y | Y | Y | Y | **Note:** Phases 3-5 operate as an evidence loop per section (retrieve → evidence store → refine outline → draft → verify claims → delta-retrieve if needed), not as strict sequential gates. --- ## Execution **On invocation, load relevant reference files:** 1. **Phase 1-7:** Load [methodology.md](./reference/methodology.md) for detailed phase instructions 2. **Phase 8 (Report):** Load [report-assembly.md](./reference/report-assembly.md) for progressive generation 3. **HTML/PDF output:** Load [html-generation.md](./reference/html-generation.md) 4. **Quality checks:** Load [quality-gates.md](./reference/quality-gates.md) 5. **Long reports (>18K words):** Load [continuation.md](./reference/continuation.md) **Templates:** - Report structure: [report_template.md](./templates/report_template.md) - HTML styling: [mckinsey_report_template.html](./templates/mckinsey_report_template.html) **Scripts:** - `python scripts/validate_report.py --report [path]` - `python scripts/verify_citations.py --report [path]` - `python scripts/md_to_html.py [markdown_path]` --- ## Output Contract **Required sections:** - Executive Summary (200-400 words) - Introduction (scope, methodology, assumptions) - Main Analysis (4-8 findings, 600-2,000 words each, cited) - Synthesis & Insights (patterns, implications) - Limitations & Caveats - Recommendations - Bibliography (COMPLETE - every citation, no placeholders) - Methodology Appendix **Output files (all to `~/Documents/[Topic]_Research_[YYYYMMDD]/`):** - Markdown (primary source of truth) - `sources.jsonl` — stable source registry with canonical IDs - `evidence.jsonl` — append-only evidence store with quotes and locators - `claims.jsonl` — atomic claim ledger with support status - `run_manifest.json` — query, mode, assumptions, provider config - HTML (McKinsey style, auto-opened) - PDF (professional print, auto-opened) **Quality standa