Research Organization Quick Reference
1-Page Summary | See Full Proposal for details
6 Research Categories
| Category | Purpose | Example Topics |
|---|---|---|
| Technology Evaluation | Assess external tools/frameworks for integration | agent-labs, copilotkit, codex |
| Academic Research | Extract patterns from academic papers | scaling-agent-systems, continual-learning |
| Competitive Intelligence | Monitor competitors and market positioning | anthropic-cowork, palantir, kimi-2.5 |
| Business & Market | Validate market opportunity and business models | product-market-fit, financial-model |
| Domain Research | Industry-specific workflows and compliance | bioscience-workorders, healthcare, regulatory |
| Process & Internal | CODITECT process improvements | system-prompt, docker-registry |
4 Lifecycle Stages
Stage 1: STAGING → analyze-new-artifacts/ (gitignored)
↓
Stage 2: ANALYSIS → internal/analysis/{category}/ (git-tracked)
↓
Stage 3: INTEGRATION → internal/architecture/ (git-tracked)
internal/research/
↓
Stage 4: ARCHIVE → internal/research/archive/ (git-tracked)
Promotion Decision Matrix
| Artifact | Staging → Analysis | Analysis → Integration | Condition |
|---|---|---|---|
| Executive Summary | ✅ Always | ❌ Never | Research complete |
| Assessment Doc | ✅ Always | ❌ Never | Quality validated |
| SDD | ⚠️ If referenced | ✅ If Go decision | ADR approved |
| TDD | ⚠️ If referenced | ✅ If Go decision | ADR approved |
| ADR | ⚠️ If referenced | ✅ When finalized | Status: Accepted/Rejected |
| C4 Diagrams | ⚠️ If referenced | ✅ If Go decision | ADR approved |
| Business Case | ✅ Always | ✅ Always | Complete |
| Financial Model | ✅ Always | ✅ Always | Validated |
| Glossary | ⚠️ If domain-specific | ✅ If reusable | 10+ terms |
| Dashboards (JSX) | ❌ Rarely | ❌ Rarely | Only if reusable |
| Source Materials | ❌ Rarely | ❌ Rarely | Only if proprietary |
Permanent Directory Structure
internal/
├── analysis/ # Stage 2: Assessments
│ ├── technology-evaluation/
│ ├── academic/
│ ├── competitive-intelligence/
│ ├── business-market/
│ ├── domain/
│ └── process-internal/
│
├── architecture/ # Stage 3: Architecture
│ ├── adrs/
│ ├── integrations/{topic}/
│ ├── domain/{industry}/
│ └── academic-patterns/
│
└── research/ # Stage 3: Research
├── business/
├── market-research/
├── domain/{industry}/
├── source-materials/
└── archive/ # Stage 4: Historical
File Naming Conventions
| File Type | Pattern | Example |
|---|---|---|
| Assessment | {topic}-assessment-YYYY-MM-DD.md | agent-labs-assessment-2026-02-16.md |
| Executive Summary | executive-summary.md | (standard name) |
| SDD | SDD-{topic}-integration.md | SDD-agent-labs-integration.md |
| TDD | TDD-{topic}-integration.md | TDD-agent-labs-integration.md |
| ADR | ADR-XXX-{topic}-{decision}.md | ADR-207-agent-labs-adoption.md |
| Business Case | {topic}-business-case-YYYY-MM-DD.md | product-market-fit-business-case-2026-02-16.md |
| Financial Model | {topic}_v{X}.{Y}.xlsx | financial-model_v3.1.xlsx |
| Manifest | MANIFEST.md | (standard name, all topics) |
Manifest Contents (Required Fields)
## Metadata
- Research Date
- Researcher
- Category
- Pipeline (Manual / ADR-206)
- Status (Active / Integrated / Archived)
## Research Objective
[What question was this research answering?]
## Source Materials
[PDFs, transcripts, GitHub repos, URLs]
## Generated Artifacts
[Phase 1: markdown, Phase 2: dashboards, Phase 3: ADRs]
## Integration Summary
- Go/No-Go decision
- Integrated components
- Remaining in staging
## Follow-up Actions
[Task IDs, implementation TODOs]
Promotion Commands (Proposed)
# Promote to analysis
/research-promote --topic "agent-labs" --to analysis
# Promote to integration (after Go decision)
/research-promote --topic "agent-labs" --to integration --artifacts sdd,tdd,adrs
# Archive superseded research
/research-archive --topic "palantir-2025" --reason "superseded by palantir-2026"
# Generate manifest
/generate-manifest --topic "agent-labs" --category "technology-evaluation"
# ADR-206 pipeline with auto-promotion
/research-pipeline "agent-labs" \
--urls https://example.com \
--github https://github.com/org/repo \
--genesis \ # Handoff to /new-project
--auto-promote # Auto-promote to analysis + integration
Migration Strategy (98 Existing Directories)
| Phase | Action | Duration |
|---|---|---|
| 1. Categorize | Auto-assign categories | 1 day |
| 2. Assess | Human review: promote/archive/delete | 5 days |
| 3. Promote | Semi-automated promotion (30 dirs) | 2 days |
| 4. Archive | Manual archive (20 dirs) | 1 day |
| 5. Delete | Grace period + auto-delete (48 dirs) | 90 days |
Result: 98 dirs → 13 active + 40 permanent (30 analysis + 15 archive) + 48 deleted
ADR-206 Integration
Pipeline Outputs (Phase 1-3):
- 9 markdown artifacts (quick-start, SDD, TDD, glossary, mermaid, ADRs, etc.)
- 4-6 JSX dashboards (tech analyzer, strategic fit, integration playbook, etc.)
- 15-25 follow-up prompts
Enhanced with This Proposal:
- ✅
MANIFEST.mdauto-generated - ✅
--auto-promoteflag for immediate promotion to analysis - ✅
--categoryauto-detection from inputs - ✅ Integration handoff (if
--genesisflag)
Key Metrics
| Metric | Before | After |
|---|---|---|
| Staging Clutter | 98 dirs | ≤13 active |
| Searchable Research | 20% | 100% |
| Artifact Loss | 80% | 0% |
| Git Repo Size | +6.1GB | +200MB |
When to Use Each Stage
Stage 1: STAGING (analyze-new-artifacts/)
- ✅ Active research in progress
- ✅ Work-in-progress artifacts
- ✅ Source materials (PDFs, transcripts)
- ✅ Generated drafts (pipelines, manual sessions)
- ❌ Anything needing permanent reference
Stage 2: ANALYSIS (internal/analysis/)
- ✅ Completed research with findings
- ✅ Executive summaries
- ✅ Assessment documents (synthesis)
- ✅ Recommendations
- ❌ SDDs/TDDs/ADRs (those go to Stage 3)
Stage 3: INTEGRATION (internal/architecture/ or internal/research/)
- ✅ Adopted SDDs/TDDs
- ✅ Finalized ADRs (Accepted/Rejected)
- ✅ Business cases
- ✅ Financial models
- ✅ Domain glossaries
- ✅ Reusable patterns
- ❌ Research-specific artifacts
Stage 4: ARCHIVE (internal/research/archive/)
- ✅ Superseded research
- ✅ Deprecated technology evaluations
- ✅ Reversed decisions
- ✅ Time-sensitive research (aged out)
- ❌ Anything currently relevant
Decision Tree
Is research complete?
NO → Stay in STAGING
YES ↓
Is assessment document written?
NO → Research incomplete, stay in STAGING
YES ↓ Promote to ANALYSIS
Is there a Go decision (ADR approved)?
NO → Stay in ANALYSIS (assessment only)
YES ↓ Promote SDDs/TDDs/ADRs to INTEGRATION
Is research time-sensitive?
NO → Stay in ANALYSIS indefinitely
YES ↓
Is research > 12 months old and not refreshed?
NO → Stay in ANALYSIS
YES ↓ ARCHIVE (with superseded_by reference)
Example: Agent Labs Research Flow
Staging (analyze-new-artifacts/agent-labs/):
- arXiv paper PDF
- Generated artifacts (SDD, TDD, 3 ADRs, 4 dashboards)
- Pipeline report
Analysis (internal/analysis/technology-evaluation/agent-labs/):
agent-labs-assessment-2026-02-16.md(human-written synthesis)executive-summary.md(promoted from staging)MANIFEST.md(lineage tracking)README.md(navigation)
Integration (internal/architecture/integrations/agent-labs/):
SDD-agent-labs-integration.mdTDD-agent-labs-integration.mdc4-diagrams/(context, container, component)
Integration (internal/architecture/adrs/):
ADR-207-agent-labs-adoption.mdADR-208-agent-labs-integration-pattern.mdADR-209-agent-orchestration-mapping.md
Staging (remains):
- arXiv paper PDF (source material)
- Glossary (topic-specific)
- Dashboards (research-specific)
Success Criteria
✅ Any contributor can find research via manifest index ✅ SDDs/TDDs/ADRs reused across projects ✅ ADR-206 pipeline outputs directly to permanent locations ✅ Clear rules for promotion/archival (no ambiguity) ✅ Manifest system self-documenting, low maintenance ✅ Zero loss of valuable artifacts ✅ 87% reduction in staging clutter
Full Proposal: research-artifact-organization-taxonomy-proposal-2026-02-16.md
Status: Proposed — 2026-02-16 Author: Claude (Sonnet 4.5)