spec-kitty-cli 0.12.1__py3-none-any.whl

specify_cli/missions/research/command-templates/tasks.md

@@ -0,0 +1,225 @@
+---
+description: Generate research work packages with subtasks aligned to methodology phases.
+---
+
+# Command Template: /spec-kitty.tasks (Research Mission)
+
+**Phase**: Design (finalizing work breakdown)
+**Purpose**: Break research work into independently executable work packages with subtasks.
+
+## User Input
+
+```text
+$ARGUMENTS
+```
+
+You **MUST** consider the user input before proceeding (if not empty).
+
+## Location Pre-flight Check
+
+Verify you are in the main repository (not a worktree). Task generation happens in main for ALL missions.
+
+```bash
+git branch --show-current  # Should show "main"
+```
+
+**Note**: Task generation in main is standard for all spec-kitty missions. Implementation happens in per-WP worktrees.
+
+---
+
+## Outline
+
+1. **Setup**: Run `spec-kitty agent feature check-prerequisites --json --paths-only --include-tasks`
+
+   **CRITICAL**: The command returns JSON with `FEATURE_DIR` as an ABSOLUTE path (e.g., `/Users/robert/Code/project/kitty-specs/015-research-topic`).
+
+   **YOU MUST USE THIS PATH** for ALL subsequent file operations.
+
+2. **Load design documents**:
+   - spec.md (research question, scope, objectives)
+   - plan.md (methodology, phases, quality criteria)
+   - research.md (background, prior art)
+   - data-model.md (entities, relationships)
+
+3. **Derive fine-grained subtasks**:
+
+   ### Subtask Patterns for Research
+
+   **Literature Search & Source Collection** (Phase 1):
+   - T001: Define search keywords and inclusion/exclusion criteria
+   - T002: [P] Search academic database 1 (IEEE, PubMed, arXiv, etc.)
+   - T003: [P] Search academic database 2
+   - T004: [P] Search gray literature and industry sources
+   - T005: Screen collected sources for relevance
+   - T006: Populate source-register.csv with all candidate sources
+   - T007: Prioritize sources by relevance rating
+
+   **Source Review & Evidence Extraction** (Phase 2):
+   - T010: [P] Review high-relevance sources (parallelizable by source)
+   - T011: Extract key findings into evidence-log.csv
+   - T012: Assign confidence levels to findings
+   - T013: Document limitations and caveats
+   - T014: Identify patterns/themes emerging from evidence
+
+   **Analysis & Synthesis** (Phase 3):
+   - T020: Code findings by theme/category
+   - T021: Identify patterns across sources and confidence levels
+   - T022: Assess strength of evidence supporting each claim
+   - T023: Draw conclusions mapped to sub-questions
+   - T024: Document limitations and threats to validity
+   - T025: Write findings.md with synthesis and bibliography references
+
+   **Quality & Validation** (Phase 4):
+   - T030: Verify source coverage meets minimum requirements
+   - T031: Validate evidence citations are traceable
+   - T032: Check for bias in source selection
+   - T033: Review methodology adherence
+   - T034: External validation (if applicable)
+
+4. **Roll subtasks into work packages**:
+
+   ### Work Package Patterns for Research
+
+   **Standard Research Flow**:
+   - WP01: Literature Search & Source Collection (T001-T007)
+   - WP02: Source Review & Evidence Extraction (T010-T014)
+   - WP03: Analysis & Synthesis (T020-T025)
+   - WP04: Quality Validation (T030-T034)
+
+   **Empirical Research (if applicable)**:
+   - WP01: Literature Review (background, prior art)
+   - WP02: Study Design & Setup
+   - WP03: Data Collection
+   - WP04: Analysis & Findings
+   - WP05: Quality Validation
+
+   **Multi-Researcher Parallel**:
+   - WP01: Search & Collect (foundation)
+   - WP02a: [P] Source Review - Researcher 1 batch
+   - WP02b: [P] Source Review - Researcher 2 batch
+   - WP03: Synthesis (depends on WP02a, WP02b)
+   - WP04: Validation
+
+   ### Prioritization
+
+   - **P0 (foundation)**: Literature search setup, source register initialization
+   - **P1 (critical)**: Source review, evidence extraction
+   - **P2 (important)**: Analysis, synthesis, findings
+   - **P3 (polish)**: Quality validation, external review
+
+5. **Write `tasks.md`**:
+   - Location: `FEATURE_DIR/tasks.md`
+   - Use `templates/tasks-template.md` from research mission
+   - Include work packages with subtasks
+   - Mark parallel opportunities (`[P]`)
+   - Define dependencies (Phase 2 depends on Phase 1, etc.)
+   - Reference evidence tracking requirements (source-register.csv, evidence-log.csv)
+
+6. **Generate prompt files**:
+
+   **CRITICAL PATH RULE**: All work package files MUST be created in a FLAT `FEATURE_DIR/tasks/` directory, NOT in subdirectories!
+
+   - Create flat `FEATURE_DIR/tasks/` directory (no subdirectories!)
+   - For each work package:
+     - Derive a kebab-case slug from the title; filename: `WPxx-slug.md`
+     - Full path: `FEATURE_DIR/tasks/WP01-literature-search.md`
+     - Use `templates/task-prompt-template.md` to capture:
+       - **YAML frontmatter with `lane: "planned"`** (CRITICAL - this is how review finds WPs!)
+       - `work_package_id`, `subtasks` array, `dependencies`, history entry
+       - Objectives, context, methodology guidance per subtask
+       - Evidence tracking requirements
+       - Quality validation criteria
+
+   **IMPORTANT**: All WP files live in flat `tasks/` directory. Lane status is tracked ONLY in the `lane:` frontmatter field, NOT by directory location. Agents can change lanes by editing the `lane:` field directly or using `spec-kitty agent tasks move-task`.
+
+7. **Finalize tasks with dependency parsing and commit**:
+
+   **CRITICAL**: Run this command from repo root:
+   ```bash
+   spec-kitty agent feature finalize-tasks --json
+   ```
+
+   This step is MANDATORY. Without it:
+   - Dependencies won't be in frontmatter
+   - Tasks won't be committed to main
+
+8. **Report**:
+   - Path to tasks.md
+   - Work package count and subtask tallies
+   - Parallelization opportunities (different sources, databases)
+   - MVP recommendation (typically WP01 Literature Search)
+   - Next command: `/spec-kitty.implement WP01`
+
+---
+
+## Research-Specific Task Generation Rules
+
+**Evidence Tracking**:
+- Every subtask that produces findings MUST specify output to evidence-log.csv
+- Every subtask that identifies sources MUST specify output to source-register.csv
+- Include subtasks for evidence validation and citation verification
+
+**Parallel Opportunities**:
+- Database searches are parallel (`[P]`) - different databases can be searched simultaneously
+- Source reviews are parallel (`[P]`) - different sources can be reviewed simultaneously
+- Researcher batches are parallel (`[P]`) - work can be split across reviewers
+
+**Quality Subtasks**:
+- Include confidence level assignment for findings
+- Include bias checking for source selection
+- Include methodology adherence verification
+
+**Work Package Scope**:
+- Each methodology phase typically gets its own work package
+- Phase transitions are natural dependency boundaries
+- Quality validation is always the final work package
+
+---
+
+## YAML Frontmatter Format (CRITICAL)
+
+**Every WP prompt file MUST use this frontmatter format**:
+
+```yaml
+---
+work_package_id: "WP01"
+subtasks:
+  - "T001"
+  - "T002"
+title: "Literature Search & Source Collection"
+phase: "Phase 1 - Literature Review"
+lane: "planned"  # DO NOT EDIT - use: spec-kitty agent tasks move-task <WPID> --to <lane>
+assignee: ""
+agent: ""
+shell_pid: ""
+review_status: ""
+reviewed_by: ""
+dependencies: []  # Added by finalize-tasks
+history:
+  - timestamp: "2026-01-19T00:00:00Z"
+    lane: "planned"
+    agent: "system"
+    shell_pid: ""
+    action: "Prompt generated via /spec-kitty.tasks"
+---
+```
+
+**DO NOT use markdown status fields like `**Status**: planned`**. The `lane:` field in YAML frontmatter is the ONLY status tracking mechanism. The review command (`/spec-kitty.review`) searches for `lane: "for_review"` in frontmatter to find WPs ready for review.
+
+---
+
+## Key Guidelines
+
+**For Agents**:
+- Use methodology phases as natural WP boundaries
+- Mark parallel subtasks (database searches, source reviews)
+- Include evidence tracking in every WP prompt
+- Quality validation depends on all content WPs
+- Use `spec-kitty agent tasks move-task` to change lanes
+
+**For Users**:
+- Tasks.md shows the full research work breakdown
+- Work packages follow methodology phases
+- MVP is typically the literature search phase
+- Parallel database searches speed up Phase 1
+- Evidence artifacts (CSV files) are tracked throughout

specify_cli/missions/research/mission.yaml

@@ -0,0 +1,115 @@
+name: "Deep Research Kitty"
+description: "Conduct systematic research with structured methodology and evidence synthesis"
+version: "1.0.0"
+domain: "research"
+
+# Workflow customization
+workflow:
+  phases:
+    - name: "question"
+      description: "Define research question and scope"
+    - name: "methodology"
+      description: "Design research methodology"
+    - name: "gather"
+      description: "Collect data and sources"
+    - name: "analyze"
+      description: "Analyze findings"
+    - name: "synthesize"
+      description: "Synthesize results and draw conclusions"
+    - name: "publish"
+      description: "Prepare findings for publication"
+
+# Expected artifacts
+artifacts:
+  required:
+    - spec.md              # Research question and scope
+    - plan.md              # Methodology plan
+    - tasks.md             # Research work packages
+    - findings.md          # Synthesized findings
+  optional:
+    - sources/             # Source documents
+    - data/                # Raw data
+    - analysis/            # Analysis outputs
+    - literature-review.md # Optional extended literature review
+    - methodology.md       # Detailed methodology appendix
+    - synthesis.md         # Synthesis notes or drafts
+
+# Path conventions for this mission
+paths:
+  workspace: "research/"
+  data: "data/"
+  deliverables: "findings/"
+  documentation: "reports/"
+
+# Validation rules
+validation:
+  checks:
+    - all_sources_documented
+    - methodology_clear
+    - findings_synthesized
+    - no_unresolved_questions
+  custom_validators: true  # Use validators.py
+
+# MCP tools recommended for this mission
+mcp_tools:
+  required:
+    - filesystem
+    - git
+  recommended:
+    - web-search
+    - pdf-reader
+    - citation-manager
+    - arxiv-search
+  optional:
+    - data-analysis
+    - statistics
+    - pubmed-search
+
+# Agent personality/instructions
+agent_context: |
+  You are a research agent conducting systematic literature reviews and empirical research.
+  Your mission is to maintain research integrity and methodological rigor throughout every phase.
+
+  Key Practices:
+  - Document ALL sources in research/source-register.csv with proper citations, URLs, and access dates.
+  - Extract findings to research/evidence-log.csv with confidence levels (high/medium/low) and contextual notes.
+  - Clearly articulate methodology decisions for reproducibility and peer review.
+  - Distinguish raw evidence from interpretation; cite every claim.
+  - Identify limitations, threats to validity, and alternative explanations.
+  - Synthesize findings so that each conclusion traces back to documented evidence rows.
+
+  Citation Standards:
+  - Use BibTeX or APA format (include DOI or URL when available).
+  - Track access dates for all online sources.
+  - Assign confidence levels commensurate with source quality.
+
+  Workflow Phases: question → methodology → gather → analyze → synthesize → publish.
+
+# Task metadata fields
+task_metadata:
+  required:
+    - task_id
+    - lane
+    - phase
+    - agent
+  optional:
+    - shell_pid
+    - assignee
+    - estimated_hours
+    - experiment_id
+    - data_collection_round
+
+# Command customization
+commands:
+  specify:
+    prompt: "Define research question, scope, and expected outcomes"
+  plan:
+    prompt: "Design research methodology and data collection strategy"
+  tasks:
+    prompt: "Break into research work packages (literature review, data collection, analysis)"
+  implement:
+    prompt: "Execute research plan and collect findings"
+  review:
+    prompt: "Review findings for validity, completeness, and proper documentation"
+  accept:
+    prompt: "Validate research completeness and readiness for publication"

specify_cli/missions/research/templates/data-model-template.md

@@ -0,0 +1,33 @@
+# Data Model (Discovery Draft)
+
+Use this skeleton to capture entities, attributes, and relationships uncovered during research. Update it as the solution space becomes clearer; implementation will refine and extend it.
+
+## Entities
+
+### Entity: <!-- e.g., Customer -->
+- **Description**: <!-- Summary -->
+- **Attributes**:
+  - `field_name` (type) – purpose / constraints
+- **Identifiers**: <!-- Primary / alternate keys -->
+- **Lifecycle Notes**: <!-- Creation, updates, archival -->
+
+### Entity: <!-- e.g., Account -->
+- **Description**:
+- **Attributes**:
+  - `field_name` (type) – purpose / constraints
+- **Identifiers**:
+- **Lifecycle Notes**:
+
+## Relationships
+
+| Source | Relation | Target | Cardinality | Notes |
+|--------|----------|--------|-------------|-------|
+| <!-- e.g., Customer --> | <!-- owns --> | <!-- Account --> | <!-- 1:N --> | <!-- Business rules --> |
+
+## Validation & Governance
+
+- **Data quality requirements**: <!-- e.g., null-handling, accepted ranges -->
+- **Compliance considerations**: <!-- PII, retention policies, regional restrictions -->
+- **Source of truth**: <!-- Systems or datasets authoritative for each entity -->
+
+> Treat this as a working model. When research uncovers new flows or systems, update the entities and relationships immediately so the implementation team inherits up-to-date context.

specify_cli/missions/research/templates/plan-template.md

@@ -0,0 +1,161 @@
+# Research Plan: [RESEARCH QUESTION]
+
+**Branch**: `[###-research-name]` | **Date**: [DATE] | **Spec**: [link]
+
+## Summary
+[One paragraph: research question + methodology + expected outcomes]
+
+## Research Context
+
+**Research Question**: [Primary question]  
+**Research Type**: Literature Review | Empirical Study | Case Study  
+**Domain**: [Academic field or industry domain]  
+**Time Frame**: [When research will be conducted]  
+**Resources Available**: [Databases, tools, budget, time]
+
+**Key Background**:
+- [Context point 1]
+- [Context point 2]
+
+## Methodology
+
+### Research Design
+
+**Approach**: [Systematic Literature Review | Survey | Experiment | Mixed Methods]
+
+**Phases**:
+1. **Question Formation** (Week 1)
+   - Define precise research question
+   - Identify sub-questions
+   - Establish scope and boundaries
+2. **Methodology Design** (Week 1-2)
+   - Select data collection methods
+   - Define analysis framework
+   - Establish quality criteria
+3. **Data Gathering** (Week 2-4)
+   - Search academic databases
+   - Screen sources for relevance
+   - Extract key findings
+   - Populate `research/evidence-log.csv`
+4. **Analysis** (Week 4-5)
+   - Code and categorize findings
+   - Identify patterns and themes
+   - Assess evidence quality
+5. **Synthesis** (Week 5-6)
+   - Draw conclusions
+   - Address research question
+   - Identify limitations
+6. **Publication** (Week 6)
+   - Write findings report
+   - Prepare presentation
+   - Share results
+
+### Data Sources
+
+**Primary Sources**:
+- [Database 1: e.g., IEEE Xplore, PubMed, arXiv]
+- [Database 2]
+
+**Secondary Sources**:
+- [Gray literature, industry reports, etc.]
+
+**Search Strategy**:
+- **Keywords**: [List search terms]
+- **Inclusion Criteria**: [What qualifies for review]
+- **Exclusion Criteria**: [What will be filtered out]
+
+### Analysis Framework
+
+**Coding Scheme**: [How findings will be categorized]  
+**Synthesis Method**: [Thematic analysis | Meta-analysis | Narrative synthesis]  
+**Quality Assessment**: [How source quality will be evaluated]
+
+## Data Management
+
+### Evidence Tracking
+
+**File**: `research/evidence-log.csv`  
+**Purpose**: Track all evidence collected with citations and findings
+
+**Columns**:
+- `timestamp`: When evidence collected (ISO format)
+- `source_type`: journal | conference | book | web | preprint
+- `citation`: Full citation (BibTeX or APA format)
+- `key_finding`: Main takeaway from this source
+- `confidence`: high | medium | low
+- `notes`: Additional context or caveats
+
+**Agent Guidance**:
+1. Read source and extract key finding.
+2. Add row to evidence-log.csv.
+3. Assign confidence level based on source quality and clarity.
+4. Note limitations or alternative interpretations.
+
+### Source Registry
+
+**File**: `research/source-register.csv`  
+**Purpose**: Maintain master list of all sources for bibliography
+
+**Columns**:
+- `source_id`: Unique identifier (e.g., "smith2025")
+- `citation`: Full citation
+- `url`: Link to source (if available)
+- `accessed_date`: When source was accessed
+- `relevance`: high | medium | low
+- `status`: reviewed | pending | archived
+
+**Agent Guidance**:
+1. Add source to register when first discovered.
+2. Update status as research progresses.
+3. Maintain relevance ratings to prioritize review.
+
+## Project Structure
+
+### Documentation (this research project)
+```
+kitty-specs/[###-research]/
+├── spec.md              # Research question and scope
+├── plan.md              # This file - methodology
+├── tasks.md             # Research work packages
+├── findings.md          # Research findings
+├── research/
+│   ├── evidence-log.csv      # All evidence with citations
+│   ├── source-register.csv   # Master source list
+│   └── methodology.md        # Detailed methodology (optional)
+└── data/                # Raw data (if empirical)
+```
+
+### Deliverables
+```
+findings/
+├── report.md           # Main research report
+├── bibliography.md     # Formatted bibliography
+└── presentation/       # Slides or summary (optional)
+```
+
+## Quality Gates
+
+### Before Data Gathering
+- [ ] Research question is clear and focused
+- [ ] Methodology is documented and reproducible
+- [ ] Data sources identified and accessible
+- [ ] Analysis framework defined
+
+### During Data Gathering
+- [ ] All sources documented in source-register.csv
+- [ ] Evidence logged with proper citations
+- [ ] Confidence levels assigned
+- [ ] Quality threshold maintained
+
+### Before Synthesis
+- [ ] All sources reviewed
+- [ ] Findings coded and categorized
+- [ ] Patterns identified
+- [ ] Limitations documented
+
+### Before Publication
+- [ ] Research question answered
+- [ ] All claims cited
+- [ ] Methodology clear and reproducible
+- [ ] Findings synthesized
+- [ ] Bibliography complete

specify_cli/missions/research/templates/research/evidence-log.csv

@@ -0,0 +1,18 @@
+# Evidence Log: Track all research findings with citations
+#
+# Column Definitions:
+# - timestamp: When evidence was collected (ISO format: YYYY-MM-DDTHH:MM:SS)
+# - source_type: journal | conference | book | web | preprint
+# - citation: Full citation (BibTeX, APA, or Simple format)
+# - key_finding: Main takeaway from this source (1-2 sentences)
+# - confidence: high | medium | low (based on source quality and clarity)
+# - notes: Additional context, caveats, or methodological notes
+#
+# Validation: Citations must be non-empty, source_type must be valid
+# Format: BibTeX (@article{...}) or APA (Author (Year). Title.) recommended
+timestamp,source_type,citation,key_finding,confidence,notes
+2025-01-15T10:00:00,journal,"Smith, J. (2024). AI Code Assistants. Nature Comp Sci, 10(2), 123-145.",AI assistants improve productivity 30%,high,Meta-analysis of 50 peer-reviewed studies
+2025-01-15T11:30:00,conference,"@inproceedings{jones2024copilot,author={Jones et al.},title={Copilot Study},booktitle={ICSE},year={2024}}",Copilot acceptance rate averages 65%,medium,GitHub internal usage study
+2025-01-15T14:00:00,web,"GitHub Copilot Stats. https://github.blog/copilot-stats-2024",35M developers actively use Copilot,medium,Self-reported marketing data
+2025-01-15T15:30:00,preprint,"Lee, K. (2024). AI Pair Programming. arXiv:2024.12345",Real-time suggestions reduce onboarding time,low,Preprint pending peer review
+2025-01-16T09:00:00,book,"Brown, A. (2023). The AI Developer. O'Reilly Media.",Context-aware tools reduce cognitive load,high,Chapter 7 includes empirical study

specify_cli/missions/research/templates/research/source-register.csv

@@ -0,0 +1,18 @@
+# Source Register: Master list of all research sources
+#
+# Column Definitions:
+# - source_id: Unique identifier (lowercase, no spaces, e.g., "smith2024")
+# - citation: Full citation (BibTeX or APA format)
+# - url: Direct link to source (DOI, arXiv, or web URL)
+# - accessed_date: Date source was accessed (YYYY-MM-DD)
+# - relevance: high | medium | low (to primary research question)
+# - status: reviewed | pending | archived
+#
+# Usage: Add sources when discovered, update status as research progresses
+# Validation: source_id must be unique
+source_id,citation,url,accessed_date,relevance,status
+smith2024,"Smith, J. (2024). AI Code Assistants. Nature Comp Sci, 10(2), 123-145.",https://doi.org/10.1038/example,2025-01-15,high,reviewed
+jones2024,"@inproceedings{jones2024copilot,author={Jones et al.},title={Copilot Study},booktitle={ICSE 2024},year={2024}}",https://dl.acm.org/example,2025-01-15,high,reviewed
+github2024,"GitHub Copilot Statistics. GitHub Blog.",https://github.blog/copilot-stats-2024,2025-01-15,medium,reviewed
+lee2024,"Lee, K. (2024). AI Pair Programming. arXiv:2024.12345",https://arxiv.org/abs/2024.12345,2025-01-15,low,pending
+brown2023,"Brown, A. (2023). The AI Developer: Human-Machine Collaboration. O'Reilly Media.",https://oreilly.com/example,2025-01-16,high,archived

specify_cli/missions/research/templates/research-template.md

@@ -0,0 +1,35 @@
+# Research Decision Log
+
+Document the outcomes of Phase 0 discovery work. Capture every clarification you resolved and the supporting evidence that backs each decision.
+
+## Summary
+
+- **Feature**: <!-- e.g., 001-pii-detection-market -->
+- **Date**: <!-- YYYY-MM-DD -->
+- **Researchers**: <!-- Primary collaborators -->
+- **Open Questions**: <!-- Items still unanswered after Phase 0 -->
+
+## Decisions & Rationale
+
+For each decision, include the supporting sources and why the team aligned on this direction.
+
+| Decision | Rationale | Evidence | Status |
+|----------|-----------|----------|--------|
+| <!-- e.g., "Adopt vector database X" --> | <!-- Why this choice was made --> | <!-- Link to evidence rows / notes --> | <!-- final / follow-up --> |
+
+## Evidence Highlights
+
+Summarize the most impactful findings from the evidence log. Link back to specific rows so the trail is auditable.
+
+- **Key insight 1** – <!-- Cite evidence row or page reference -->
+- **Key insight 2** – <!-- Cite evidence row or page reference -->
+- **Risks / Concerns** – <!-- Any red flags uncovered -->
+
+## Next Actions
+
+Outline what needs to happen before moving into implementation planning.
+
+1. <!-- e.g., Validate data access with security team -->
+2. <!-- e.g., Schedule stakeholder read-out -->
+
+> Keep this document living. As more evidence arrives, update decisions and rationale so downstream implementers can trust the history.

specify_cli/missions/research/templates/spec-template.md

@@ -0,0 +1,64 @@
+# Research Specification: [RESEARCH QUESTION]
+
+**Feature Branch**: `[###-research-name]`  
+**Created**: [DATE]  
+**Status**: Draft  
+**Research Type**: Literature Review | Empirical Study | Case Study | Meta-Analysis
+
+## Research Question & Scope
+
+**Primary Research Question**: [What specific question does this research aim to answer?]
+
+**Sub-Questions**:
+1. [Supporting question 1]
+2. [Supporting question 2]
+
+**Scope**:
+- **In Scope**: [What will be investigated]
+- **Out of Scope**: [What will NOT be investigated]
+- **Boundaries**: [Time period, geographic region, specific domains]
+
+**Expected Outcomes**:
+- [What findings or artifacts will this research produce?]
+- [How will results be applied?]
+
+## Research Methodology Outline
+
+### Research Approach
+- **Method**: Systematic Literature Review | Survey | Experiment | Case Study
+- **Data Sources**: [Where will data / sources come from?]
+- **Analysis Approach**: [How will data be analyzed?]
+
+### Success Criteria
+- [Measurable criterion 1: e.g., "Review at least 50 peer-reviewed sources"]
+- [Measurable criterion 2: e.g., "Identify 3+ validated patterns"]
+- [Quality criterion: e.g., "All sources cited in APA/BibTeX format"]
+
+## Research Requirements
+
+### Data Collection Requirements
+- **DR-001**: Research MUST collect data from [specific sources]
+- **DR-002**: All sources MUST be documented in `research/source-register.csv`
+- **DR-003**: Citations MUST follow [BibTeX | APA | other] format
+
+### Analysis Requirements
+- **AR-001**: Findings MUST be synthesized into [specific output]
+- **AR-002**: Methodology MUST be clearly documented and reproducible
+- **AR-003**: Limitations MUST be explicitly identified
+
+### Quality Requirements
+- **QR-001**: All claims MUST be supported by cited evidence
+- **QR-002**: Confidence levels MUST be assigned to findings in `research/evidence-log.csv`
+- **QR-003**: Alternative interpretations MUST be considered
+
+## Key Concepts & Terminology
+
+- **[Concept 1]**: [Definition in research context]
+- **[Concept 2]**: [Definition relevant to research question]
+- **[Concept 3]**: [Any mission-specific terminology]
+
+## Evidence Tracking Guidance
+
+- Log every reviewed source in `research/source-register.csv` with citation, URL, relevance, and status.
+- Capture each key finding in `research/evidence-log.csv`, including confidence level and notes.
+- Reference evidence row IDs within this specification when making claims.