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:
- Phase 1-7: Load methodology.md for detailed phase instructions
- Phase 8 (Report): Load report-assembly.md for progressive generation
- HTML/PDF output: Load html-generation.md
- Quality checks: Load quality-gates.md
- Long reports (>18K words): Load continuation.md
Templates:
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 standards:
- 10+ sources, 3+ per major claim (cluster-independent, not just count)
- All factual claims cited immediately [N] with evidence backing in
evidence.jsonl
- Claim-support verification mandatory: no unsupported factual claims pass delivery
- No placeholders, no fabricated citations
- Prose-first (>=80%), bullets sparingly
When to Use / NOT Use
Use: Comprehensive analysis, technology comparisons, state-of-the-art reviews, multi-perspective investigation, market analysis.
Do NOT use: Simple lookups, debugging, 1-2 search answers, quick time-sensitive queries.
1---
2name: deep-research
3description: 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.
4metadata:
5 source: https://skillmd.com/skills/199-biotechnologies/deep-research
6---
7
8# Deep Research
9
10## Core Purpose
11
12Deliver citation-tracked research reports through a structured pipeline with evidence persistence, source identity management, claim-level verification, and progressive context management.
13
14**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.
15
16---
17
18## Decision Tree
19
20```
21Request Analysis
22+-- Simple lookup? --> STOP: Use WebSearch
23+-- Debugging? --> STOP: Use standard tools
24+-- Complex analysis needed? --> CONTINUE
25
26Mode Selection
27+-- Initial exploration --> quick (3 phases, 2-5 min)
28+-- Standard research --> standard (6 phases, 5-10 min) [DEFAULT]
29+-- Critical decision --> deep (8 phases, 10-20 min)
30+-- Comprehensive review --> ultradeep (8+ phases, 20-45 min)
31```
32
33**Default assumptions:** Technical query = technical audience. Comparison = balanced perspective. Trend = recent 1-2 years.
34
35---
36
37## Workflow Overview
38
39| Phase | Name | Quick | Std | Deep | Ultra |
40|-------|------|-------|-----|------|-------|
41| 1 | SCOPE | Y | Y | Y | Y |
42| 2 | PLAN | - | Y | Y | Y |
43| 3 | RETRIEVE | Y | Y | Y | Y |
44| 4 | TRIANGULATE | - | Y | Y | Y |
45| 4.5 | OUTLINE REFINEMENT | - | Y | Y | Y |
46| 5 | SYNTHESIZE | - | Y | Y | Y |
47| 6 | CRITIQUE | - | - | Y | Y |
48| 7 | REFINE | - | - | Y | Y |
49| 8 | PACKAGE | Y | Y | Y | Y |
50
51**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.
52
53---
54
55## Execution
56
57**On invocation, load relevant reference files:**
58
591. **Phase 1-7:** Load [methodology.md](./reference/methodology.md) for detailed phase instructions
602. **Phase 8 (Report):** Load [report-assembly.md](./reference/report-assembly.md) for progressive generation
613. **HTML/PDF output:** Load [html-generation.md](./reference/html-generation.md)
624. **Quality checks:** Load [quality-gates.md](./reference/quality-gates.md)
635. **Long reports (>18K words):** Load [continuation.md](./reference/continuation.md)
64
65**Templates:**
66- Report structure: [report_template.md](./templates/report_template.md)
67- HTML styling: [mckinsey_report_template.html](./templates/mckinsey_report_template.html)
68
69**Scripts:**
70- `python scripts/validate_report.py --report [path]`
71- `python scripts/verify_citations.py --report [path]`
72- `python scripts/md_to_html.py [markdown_path]`
73
74---
75
76## Output Contract
77
78**Required sections:**
79- Executive Summary (200-400 words)
80- Introduction (scope, methodology, assumptions)
81- Main Analysis (4-8 findings, 600-2,000 words each, cited)
82- Synthesis & Insights (patterns, implications)
83- Limitations & Caveats
84- Recommendations
85- Bibliography (COMPLETE - every citation, no placeholders)
86- Methodology Appendix
87
88**Output files (all to `~/Documents/[Topic]_Research_[YYYYMMDD]/`):**
89- Markdown (primary source of truth)
90- `sources.jsonl` — stable source registry with canonical IDs
91- `evidence.jsonl` — append-only evidence store with quotes and locators
92- `claims.jsonl` — atomic claim ledger with support status
93- `run_manifest.json` — query, mode, assumptions, provider config
94- HTML (McKinsey style, auto-opened)
95- PDF (professional print, auto-opened)
96
97**Quality standards:**
98- 10+ sources, 3+ per major claim (cluster-independent, not just count)
99- All factual claims cited immediately [N] with evidence backing in `evidence.jsonl`
100- Claim-support verification mandatory: no unsupported factual claims pass delivery
101- No placeholders, no fabricated citations
102- Prose-first (>=80%), bullets sparingly
103
104---
105
106## When to Use / NOT Use
107
108**Use:** Comprehensive analysis, technology comparisons, state-of-the-art reviews, multi-perspective investigation, market analysis.
109
110**Do NOT use:** Simple lookups, debugging, 1-2 search answers, quick time-sensitive queries.