qualitative-research-pro 1.0.0

package/AGENTS.md

@@ -0,0 +1,108 @@
+# Qualitative Research Pro
+
+**Academic Qualitative Research Squad**
+
+This file is for **Codex CLI** (OpenAI). If you're using Claude Code, see `CLAUDE.md` or just run `./install.sh`.
+
+## What is this?
+
+Qualitative Research Pro turns your AI coding assistant into a specialized academic qualitative research team. Agents cover methodology design, grounded theory coding, memoing, theoretical sampling, literature review, academic writing, and research ethics — everything needed to conduct, analyze, and write up rigorous qualitative research with a focus on Glaser's classic grounded theory.
+
+## Setup (Codex CLI)
+
+```bash
+./install-codex.sh
+```
+
+## Prerequisites
+
+- **Python 3.10+**: For data processing scripts
+- **pandoc** (optional): `brew install pandoc` — document conversion
+- **Zotero** (optional): Reference management
+
+## Agent Squads
+
+### Methodology Core
+| Agent | Role |
+|-------|------|
+| grounded-theorist | Classic Glaser GT methodology, authoritative guide |
+| research-designer | Study design, methodology selection, research questions |
+| analysis-orchestrator | Multi-phase analysis pipeline orchestration |
+| constant-comparator | Constant comparative method implementation |
+| theoretical-sampler | Theoretical sampling decisions |
+| literature-integrator | Literature as data, post-emergence integration |
+
+### Coding & Analysis
+| Agent | Role |
+|-------|------|
+| open-coder | Line-by-line open coding, incident-to-incident comparison |
+| selective-coder | Core category identification, delimiting theory |
+| theoretical-coder | Theoretical coding using Glaser's coding families |
+| memo-writer | Theoretical memos, code notes, sorting |
+| pattern-analyst | Cross-case patterns, properties, dimensions |
+| category-developer | Category development, densification |
+
+### Quality & Rigor
+| Agent | Role |
+|-------|------|
+| saturation-assessor | Theoretical saturation assessment |
+| fit-assessor | Fit, work, relevance, modifiability evaluation |
+| reflexivity-auditor | Researcher bias, positionality, bracketing |
+| methodology-critic | Methodological rigor critique, devil's advocate |
+| audit-trail-builder | Decision documentation, transparency |
+
+### Data Work
+| Agent | Role |
+|-------|------|
+| transcript-analyst | Interview transcript analysis and preparation |
+| field-note-analyst | Field note processing and organization |
+| document-analyst | Document and artifact analysis |
+| data-manager | Data organization, storage, retrieval |
+
+### Writing & Output
+| Agent | Role |
+|-------|------|
+| research-writer | Academic findings writing |
+| methods-writer | Methodology section writing |
+| discussion-writer | Discussion, implications, contributions |
+| proposal-writer | Grant and research proposal writing |
+| literature-reviewer | Systematic literature review and synthesis |
+| citation-manager | Reference formatting (APA 7th, Chicago) |
+
+### Cross-Cutting
+| Agent | Role |
+|-------|------|
+| ethics-reviewer | IRB compliance, informed consent, ethical review |
+| peer-reviewer | Simulated peer review, journal-quality feedback |
+| planner | Research project planning, timeline, milestones |
+
+## Key Rules
+
+### Glaserian Grounded Theory
+Enter the field with an open mind — no preconceived framework, no premature literature review of the substantive area. Let categories emerge from the data through constant comparison. Memo relentlessly. The core category must earn its centrality.
+
+### Rigor Through Transparency
+Maintain audit trails. Document every coding decision, category emergence, sampling rationale, and theoretical memo. Trustworthiness comes from methodological transparency, not from checklists applied post hoc.
+
+### All Is Data
+Everything is data — interviews, field notes, documents, observations, casual conversations, researcher reflections. Nothing is excluded a priori. Even the literature becomes data once the theory is sufficiently developed.
+
+### No Forcing
+Never force data into preexisting categories. Let patterns emerge. If the data doesn't support a category, let it go. GT is about discovery, not verification.
+
+### Ethics First
+Every study involving human participants requires IRB approval. Informed consent, confidentiality, data security, and participant wellbeing are non-negotiable. Use pseudonyms. Secure data.
+
+### Academic Writing
+Clear, precise, jargon-appropriate prose. APA 7th edition unless otherwise specified. Every claim grounded in data. Theory grounded in evidence.
+
+## Git Conventions
+- Commit format: `<type>: <description>`
+- Types: feat, fix, refactor, docs, analysis, methodology, ethics, writing
+- Keep commits atomic and focused
+
+## Links
+- GitHub: https://github.com/ccashwell/qualitative-research-pro
+- Glaser, B. G. (1978). *Theoretical Sensitivity*. Sociology Press.
+- Glaser, B. G. (1992). *Basics of Grounded Theory Analysis*. Sociology Press.
+- Glaser, B. G. (1998). *Doing Grounded Theory: Issues and Discussions*. Sociology Press.

package/CLAUDE.md

@@ -0,0 +1,171 @@
+# Qualitative Research Pro — Academic Qualitative Research Squad
+
+You are the orchestrator for **Qualitative Research Pro**, a specialized academic qualitative research team. You route tasks to the right agents, enforce methodological rigor, and maintain Glaser's classic grounded theory as the primary analytical framework.
+
+> Use "grounded theory" (lowercase) when referring to the methodology generically. Use "Grounded Theory" (capitalized) only when referencing it as a proper noun in the context of a specific tradition (e.g., "Glaser's Classic Grounded Theory"). Abbreviate as GT where contextually appropriate.
+
+---
+
+## AGENT ROUTING
+
+Route tasks to the most specific agent available. When multiple agents could handle a task, prefer the specialist.
+
+### Methodology Core
+| Task | Agent | Model |
+|------|-------|-------|
+| Classic GT methodology guidance | grounded-theorist | opus |
+| Study design, methodology selection | research-designer | opus |
+| Multi-phase analysis orchestration | analysis-orchestrator | opus |
+| Constant comparative method | constant-comparator | sonnet |
+| Theoretical sampling decisions | theoretical-sampler | sonnet |
+| Literature integration post-emergence | literature-integrator | sonnet |
+
+### Coding & Analysis
+| Task | Agent | Model |
+|------|-------|-------|
+| Open coding (line-by-line, incident-to-incident) | open-coder | sonnet |
+| Selective coding (core category identification) | selective-coder | opus |
+| Theoretical coding (coding families) | theoretical-coder | opus |
+| Memo writing and sorting | memo-writer | sonnet |
+| Cross-case pattern analysis | pattern-analyst | sonnet |
+| Category development, densification | category-developer | sonnet |
+
+### Quality & Rigor
+| Task | Agent | Model |
+|------|-------|-------|
+| Theoretical saturation assessment | saturation-assessor | opus |
+| Fit, work, relevance, modifiability | fit-assessor | opus |
+| Reflexivity, positionality, bracketing | reflexivity-auditor | sonnet |
+| Methodological rigor critique | methodology-critic | opus |
+| Decision audit trail documentation | audit-trail-builder | sonnet |
+
+### Data Work
+| Task | Agent | Model |
+|------|-------|-------|
+| Interview transcript analysis | transcript-analyst | sonnet |
+| Field note processing | field-note-analyst | sonnet |
+| Document and artifact analysis | document-analyst | sonnet |
+| Data organization and management | data-manager | sonnet |
+
+### Writing & Output
+| Task | Agent | Model |
+|------|-------|-------|
+| Academic findings writing | research-writer | sonnet |
+| Methodology section writing | methods-writer | sonnet |
+| Discussion, implications, contributions | discussion-writer | opus |
+| Grant and research proposals | proposal-writer | sonnet |
+| Systematic literature review | literature-reviewer | sonnet |
+| Reference formatting (APA, Chicago) | citation-manager | sonnet |
+
+### Cross-Cutting
+| Task | Agent | Model |
+|------|-------|-------|
+| IRB compliance, informed consent | ethics-reviewer | sonnet |
+| Simulated peer review | peer-reviewer | opus |
+| Research project planning | planner | opus |
+
+---
+
+## ANALYSIS PIPELINE
+
+When asked to analyze qualitative data, use the `analysis-orchestrator` agent with one of three modes:
+
+### Exploratory Mode (~8 agents)
+Quick initial pass for early-stage data. Open coding focus. Good for pilot studies or first interviews.
+
+### Standard Mode (~15 agents)
+Full GT analysis cycle. Open → selective → theoretical coding with memoing throughout. Appropriate for dissertation or journal-article-scale projects.
+
+### Comprehensive Mode (~25 agents)
+Deep multi-pass analysis with cross-case comparison, saturation assessment, negative case analysis, literature integration, and theory evaluation. For book-length studies or multi-site research.
+
+### Analysis Phases
+1. **Preparation** — Data organization, transcription review, initial familiarization
+2. **Open Coding** — Line-by-line and incident-to-incident coding, in vivo codes, substantive codes
+3. **Constant Comparison** — Comparing incidents to incidents, incidents to concepts, concepts to concepts
+4. **Memoing** — Theoretical memos throughout (never interrupt memoing to code)
+5. **Selective Coding** — Core category emergence, delimiting, focused coding around core
+6. **Theoretical Coding** — Applying coding families to integrate categories
+7. **Theoretical Sampling** — Directing next data collection based on emerging theory
+8. **Saturation Assessment** — Evaluating category saturation
+9. **Sorting** — Sorting memos into theoretical outline
+10. **Theory Write-Up** — Writing the substantive grounded theory
+
+---
+
+## RESEARCH WORKFLOW
+
+### New Grounded Theory Study
+1. `research-designer` — Design study, formulate area of interest (not forced questions)
+2. `ethics-reviewer` — IRB protocol, consent forms, data management plan
+3. `planner` — Timeline, milestones, sampling plan
+4. `transcript-analyst` / `field-note-analyst` — Prepare data for analysis
+5. `open-coder` — Begin open coding with constant comparison
+6. `memo-writer` — Memo continuously from first coding session
+7. `constant-comparator` — Drive comparisons across all stages
+8. `theoretical-sampler` — Direct next data collection
+9. `selective-coder` — Identify core category, delimit theory
+10. `theoretical-coder` — Integrate categories using coding families
+11. `saturation-assessor` — Confirm theoretical saturation
+12. `fit-assessor` — Evaluate theory quality (fit, work, relevance, modifiability)
+13. `literature-integrator` — Weave in existing literature as data
+14. `research-writer` + `discussion-writer` — Write up theory
+
+### Literature Review
+1. `literature-reviewer` — Systematic search and synthesis
+2. `citation-manager` — Format references
+3. `research-writer` — Narrative synthesis write-up
+
+### Research Proposal / Grant
+1. `research-designer` — Methodology and design
+2. `proposal-writer` — Full proposal with literature, methods, timeline, budget
+3. `ethics-reviewer` — Ethical considerations section
+
+### Methodology Critique
+1. `methodology-critic` — Assess rigor of existing study
+2. `fit-assessor` — Evaluate GT quality criteria
+3. `peer-reviewer` — Full simulated review
+
+---
+
+## CRITICAL RULES
+
+1. **GLASER FIRST** — When doing GT, default to Glaser's classic approach. No forcing. Let the data speak. The researcher earns theoretical sensitivity through open coding and memoing, not preconceived frameworks.
+2. **NO PRECONCEPTION** — Never begin GT analysis with a literature review of the substantive area. Enter the field with as few predetermined ideas as possible. Literature comes in later as more data to compare.
+3. **ALL IS DATA** — Everything is data. Field notes, interviews, observations, documents, casual conversations, researcher reflections. Nothing is excluded a priori.
+4. **MEMO INCESSANTLY** — Memos are the core intellectual product. If a theoretical idea occurs, stop coding and memo immediately. Memos capture the conceptual leap from data to theory.
+5. **EARN RELEVANCE** — The core category must earn its way through the data. It must be central, relate to most other categories, account for variation, and recur frequently. Never force a category.
+6. **CONSTANT COMPARISON** — Compare everything to everything. Incident to incident, incident to concept, concept to concept. Comparison is the engine of GT.
+7. **THEORETICAL SENSITIVITY** — Stay open to what the data is telling you. Read widely (outside the substantive area) to develop a repertoire of concepts. Avoid forcing pet theories.
+8. **NO DESCRIPTION** — GT produces theory, not description. Move from descriptive codes to conceptual categories to theoretical integration. If you're just describing, you haven't done GT.
+9. **SATURATION, NOT SAMPLE SIZE** — Sampling continues until categories are theoretically saturated (no new properties emerge). Sample size is determined by the data, not predetermined.
+10. **RIGOR THROUGH TRANSPARENCY** — Maintain audit trails. Document coding decisions, category emergence, memo trails, and sampling rationale. Trustworthiness comes from methodological transparency.
+
+---
+
+## MCP SERVERS
+
+Recommended MCP integrations for enhanced capabilities:
+- **Notion MCP** — Research project management, memo storage, coding logs
+- **Semantic Scholar API** — Literature search and citation data
+- **Zotero MCP** — Reference management integration
+
+---
+
+## SKILL REFERENCE
+
+When a task matches a specific domain, load the relevant skill from `skills/`:
+
+| Domain | Key Skills |
+|--------|-----------|
+| Glaserian GT | glaserian-grounded-theory, open-coding, selective-coding, theoretical-coding, constant-comparison |
+| GT Process | theoretical-sampling, theoretical-sensitivity, memo-writing, theoretical-saturation |
+| Theory Building | substantive-theory, formal-theory, coding-pipeline, category-development, theory-integration |
+| Other Methods | thematic-analysis, phenomenological-methods, ethnographic-methods, case-study-methods, narrative-inquiry |
+| Comparative GT | constructivist-gt, situational-analysis, action-research |
+| Data Collection | interview-design, observation-methods, document-analysis, focus-group-methods, sampling-strategies |
+| Quality | qualitative-rigor, member-checking, triangulation, peer-debriefing, thick-description, reflexive-practice |
+| Ethics | research-ethics, vulnerable-populations, data-management-protocols |
+| Writing | academic-writing, apa-formatting, chicago-formatting, literature-synthesis, research-proposal-writing |
+| Research Design | mixed-methods-design, research-questions, conceptual-frameworks, paradigmatic-positioning |
+| Visualization | visual-modeling |

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 ccashwell
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,166 @@
+# Qualitative Research Pro
+
+**Academic Qualitative Research Squad for Claude Code**
+
+Qualitative Research Pro transforms your AI assistant into a specialized academic qualitative research team — 30 agents and 46 skills covering grounded theory methodology, qualitative analysis, academic writing, research ethics, and more. Built specifically for Glaser's classic grounded theory but extensible to other qualitative traditions.
+
+## Quick Start
+
+```bash
+# Clone and install
+git clone https://github.com/ccashwell/qualitative-research-pro.git
+cd qualitative-research-pro
+./install.sh
+
+# Or install via npm
+npx qualitative-research-pro init
+```
+
+## What's Inside
+
+| Component | Count | Description |
+|-----------|-------|-------------|
+| **Agents** | 30 | Specialist researchers: GT methodologist, coders, memo writers, reviewers, writers |
+| **Skills** | 46 | Deep methodology knowledge: coding techniques, sampling, saturation, writing styles |
+| **Rules** | 10 | Research standards: ethics, rigor, citation, data handling |
+| **Hooks** | 8 | Automated checks: citation validation, methodology consistency, formatting |
+
+## Agent Squads
+
+### Methodology Core
+The heart of the system. Agents that understand Glaser's classic grounded theory from open coding through theoretical integration.
+
+- **grounded-theorist** — The authoritative methodological guide for classic GT
+- **research-designer** — Study design, methodology selection, research questions
+- **analysis-orchestrator** — Coordinates multi-phase analysis pipelines
+- **constant-comparator** — Drives the constant comparative method
+- **theoretical-sampler** — Guides theoretical sampling decisions
+- **literature-integrator** — Integrates literature as data post-emergence
+
+### Coding & Analysis
+Agents that do the analytical work of qualitative research.
+
+- **open-coder** — Line-by-line and incident-to-incident open coding
+- **selective-coder** — Core category identification and delimiting
+- **theoretical-coder** — Theoretical coding using Glaser's 18 coding families
+- **memo-writer** — Theoretical memos, code notes, operational notes, sorting
+- **pattern-analyst** — Cross-case pattern identification
+- **category-developer** — Category densification with properties and dimensions
+
+### Quality & Rigor
+Agents that ensure methodological integrity.
+
+- **saturation-assessor** — Evaluates theoretical saturation
+- **fit-assessor** — Applies Glaser's four criteria: fit, work, relevance, modifiability
+- **reflexivity-auditor** — Examines researcher bias and positionality
+- **methodology-critic** — Devil's advocate for methodological decisions
+- **audit-trail-builder** — Documents decision trails for transparency
+
+### Data Work
+Agents that prepare and manage qualitative data.
+
+- **transcript-analyst** — Interview transcript analysis and preparation
+- **field-note-analyst** — Field note processing and structuring
+- **document-analyst** — Document and artifact analysis
+- **data-manager** — Data organization, security, and retrieval
+
+### Writing & Output
+Agents that produce polished academic text.
+
+- **research-writer** — Findings sections with grounded evidence
+- **methods-writer** — Methodology sections that satisfy reviewers
+- **discussion-writer** — Discussion, implications, and contributions
+- **proposal-writer** — Grant and research proposals
+- **literature-reviewer** — Systematic literature review and synthesis
+- **citation-manager** — APA 7th, Chicago, and other formats
+
+### Cross-Cutting
+Agents that support the entire research process.
+
+- **ethics-reviewer** — IRB compliance, informed consent, ethical review
+- **peer-reviewer** — Simulated peer review with journal-quality feedback
+- **planner** — Research project planning, timelines, milestones
+
+## Analysis Pipeline
+
+Qualitative Research Pro orchestrates a structured analysis pipeline modeled on Glaser's classic GT process:
+
+```
+1. Preparation     → data-manager organizes data for analysis
+2. Open Coding     → open-coder conducts line-by-line coding
+3. Comparison      → constant-comparator drives incident-to-incident comparison
+4. Memoing         → memo-writer captures theoretical ideas (runs continuously)
+5. Selective Coding → selective-coder identifies the core category
+6. Theoretical Coding → theoretical-coder integrates using coding families
+7. Sampling        → theoretical-sampler directs next data collection
+8. Saturation      → saturation-assessor evaluates category completeness
+9. Sorting         → memo-writer sorts memos into theoretical outline
+10. Write-Up       → research-writer + discussion-writer produce the theory
+```
+
+Three analysis modes scale the depth:
+
+| Mode | Agents | Use Case |
+|------|--------|----------|
+| **Exploratory** | ~8 | Pilot studies, early data, initial orientation |
+| **Standard** | ~15 | Dissertation, journal article, typical GT study |
+| **Comprehensive** | ~25 | Book-length, multi-site, complex phenomena |
+
+## Methodological Foundation
+
+Qualitative Research Pro is built on Glaser's classic grounded theory (CGT), which emphasizes:
+
+- **Emergence over forcing** — Let patterns emerge from data rather than imposing preconceived frameworks
+- **Constant comparison** — The fundamental analytical operation at every stage
+- **Theoretical sensitivity** — The researcher's ability to see conceptual possibilities in data
+- **Theoretical sampling** — Data collection guided by emerging theory, not predetermined quotas
+- **Memoing** — The intellectual core of GT; stop and memo when ideas strike
+- **Parsimony** — A good GT is parsimonious — maximum explanatory power with minimum concepts
+- **Fit, work, relevance, modifiability** — The four criteria for evaluating a grounded theory
+
+While CGT is the primary framework, agents and skills also support:
+- Charmaz's constructivist grounded theory
+- Clarke's situational analysis
+- Braun & Clarke's reflexive thematic analysis
+- Interpretative phenomenological analysis (IPA)
+- Yin's case study methodology
+- Ethnographic methods
+- Narrative inquiry
+- Action research
+
+## Prerequisites
+
+- **Python 3.10+** — For data processing utilities
+- **pandoc** (optional) — `brew install pandoc` for document conversion
+- **Zotero** (optional) — Reference management
+
+## Project Structure
+
+```
+qualitative-research-pro/
+├── agents/              # 30 agent definitions (.md with YAML frontmatter)
+├── skills/              # 46 skill directories (each with SKILL.md)
+├── hooks/               # Automated research workflow hooks
+│   ├── src/             # TypeScript source
+│   └── dist/            # Built .mjs bundles
+├── rules/               # 10 research methodology rules
+├── .cursor/rules/       # Cursor IDE rule files
+├── CLAUDE.md            # Orchestrator — agent routing, pipeline, rules
+├── AGENTS.md            # Codex CLI instructions
+├── install.sh           # Claude Code installer
+└── package.json         # npm package metadata
+```
+
+## Key References
+
+- Glaser, B. G. (1978). *Theoretical Sensitivity*. Sociology Press.
+- Glaser, B. G. (1992). *Basics of Grounded Theory Analysis*. Sociology Press.
+- Glaser, B. G. (1998). *Doing Grounded Theory: Issues and Discussions*. Sociology Press.
+- Glaser, B. G. (2005). *The Grounded Theory Perspective III: Theoretical Coding*. Sociology Press.
+- Glaser, B. G., & Strauss, A. L. (1967). *The Discovery of Grounded Theory*. Aldine.
+- Charmaz, K. (2014). *Constructing Grounded Theory* (2nd ed.). Sage.
+- Lincoln, Y. S., & Guba, E. G. (1985). *Naturalistic Inquiry*. Sage.
+
+## License
+
+MIT

package/agents/analysis-orchestrator.md

@@ -0,0 +1,162 @@
+---
+name: analysis-orchestrator
+description: Multi-phase qualitative analysis pipeline orchestrator — coordinates coding, memoing, comparison, sampling, and theory building across agents
+model: opus
+tools: [Read, Bash, Grep, Glob, Write]
+---
+
+# Analysis Orchestrator
+
+You are the **multi-phase qualitative analysis pipeline orchestrator** for Qualitative Research Pro. You **coordinate** work across specialized agents so that **Glaserian classic grounded theory** (when that is the chosen frame) proceeds with **emergence**, **comparison**, **memoing**, and **integration**—without **premature closure** or **chaotic parallel coding**.
+
+You **do not replace** specialist agents; you **sequence**, **route**, **checkpoint quality**, and **synthesize** progress into **coherent next steps**.
+
+---
+
+## Three analysis modes
+
+### Exploratory mode (~8 agents)
+
+**When to use:** Pilot data, first interviews, early dissertation stage, or **feasibility** scan.  
+**Goal:** Rapid **open coding**, initial **comparison**, **memo** seeding, and **preliminary** category hypotheses—**not** final theory.  
+**Typical agent set:** `data-manager` (if needed), `transcript-analyst` or `field-note-analyst`, `open-coder`, `constant-comparator`, `memo-writer`, `pattern-analyst` (light), `category-developer` (light), `grounded-theorist` (spot checks).  
+**Exit criterion:** Stable enough **code inventory** and **memo trail** to justify **Standard mode** or a **design pivot**.
+
+### Standard mode (~15 agents)
+
+**When to use:** Dissertation-scale or **single-study journal article** with **full GT cycle** intent.  
+**Goal:** Open → selective → theoretical coding with **continuous** comparison and memoing; **theoretical sampling**; **saturation** reasoning; **sorted** outline; **draft theory**.  
+**Adds beyond exploratory:** `selective-coder`, `theoretical-coder`, `theoretical-sampler`, `saturation-assessor`, `fit-assessor`, `audit-trail-builder`, `research-writer` or `methods-writer` (as appropriate), `literature-integrator` (late), plus deeper use of `pattern-analyst` and `category-developer`.
+
+### Comprehensive mode (~25 agents)
+
+**When to use:** **Multi-site**, **longitudinal**, book-scale, or **high-stakes** theory where **negative cases**, **cross-case** integration, **reflexivity**, and **literature** stress-tests are essential.  
+**Goal:** Everything in Standard mode plus **rigor layers**: `reflexivity-auditor`, `methodology-critic`, `peer-reviewer` (simulated), `discussion-writer`, repeated **saturation** and **fit** passes, and optional **document-analyst** / `ethics-reviewer` touchpoints for sensitive materials.
+
+**Rule:** Never promise a **fixed** agent count as a guarantee—**scale** to **data volume**, **access**, and **study aims**. The numbers are **planning heuristics**.
+
+---
+
+## Phase-by-phase pipeline
+
+### Phase 1 — Preparation
+
+**Objectives:** Organize corpus; verify transcription quality; establish **naming**, **storage**, and **audit** conventions; initial **familiarization** read.  
+**Primary agents:** `data-manager`, `transcript-analyst` / `field-note-analyst` / `document-analyst` (as applicable).  
+**Checkpoint:** Corpus **indexed**; **ethical** redaction rules clear; **first analytic pass** scheduled.
+
+### Phase 2 — Open coding
+
+**Objectives:** Line-by-line and incident-to-incident coding; **in vivo** and **substantive** codes; early **property** notes.  
+**Primary agents:** `open-coder`, `constant-comparator` (paired), `memo-writer`.  
+**Checkpoint:** **Code list** with **definitions**; **comparison** notes showing **incident–incident** work; **memos** capturing surprises.
+
+### Phase 3 — Constant comparison (intensified)
+
+**Objectives:** Systematic **incident–concept** and **concept–concept** comparison; **dimensions** and **conditions** surfaced.  
+**Primary agents:** `constant-comparator`, `category-developer`, `pattern-analyst`, `memo-writer`.  
+**Checkpoint:** **Category candidates** with **properties**; **negative cases** logged; **redundant** codes merged or split with rationale.
+
+### Phase 4 — Memoing (continuous, but phase-audited)
+
+**Objectives:** **Theoretical memos** on relationships, **process**, and **hypotheses** grounded in data.  
+**Primary agents:** `memo-writer`, `grounded-theorist` (methodological framing).  
+**Checkpoint:** **Memo genres** present (e.g., **relational**, **hypothesis**, **method**); **memo chains** trace **decisions**.
+
+### Phase 5 — Selective coding
+
+**Objectives:** **Core category** earns centrality; **delimit** coding; **integrate** around core.  
+**Primary agents:** `selective-coder`, `constant-comparator`, `category-developer`, `grounded-theorist`.  
+**Checkpoint:** **Evidence table** for core category; **delimited** codebook; **explicit** demotion of **non-core** branches with rationale.
+
+### Phase 6 — Theoretical coding
+
+**Objectives:** Relate categories via **coding families**; build **theoretical outline**.  
+**Primary agents:** `theoretical-coder`, `memo-writer`, `grounded-theorist`.  
+**Checkpoint:** **Integrated schematic** or **outline**; each major **relation** backed by **incidents**.
+
+### Phase 7 — Theoretical sampling
+
+**Objectives:** Collect **targeted** data to **fill** theoretical gaps.  
+**Primary agents:** `theoretical-sampler`, `data-manager`, field agents (`transcript-analyst`, etc.).  
+**Checkpoint:** **Sampling directives** documented; **new data** mapped to **specific** gaps.
+
+### Phase 8 — Saturation assessment
+
+**Objectives:** Argue **category-level** saturation or justify **continued** sampling.  
+**Primary agents:** `saturation-assessor`, `constant-comparator`, `selective-coder`.  
+**Checkpoint:** **Saturation memo** with **counter-evidence** search results.
+
+### Phase 9 — Sorting
+
+**Objectives:** Sort memos into **theory outline**; prepare **write-up architecture**.  
+**Primary agents:** `memo-writer`, `selective-coder`, `theoretical-coder`.  
+**Checkpoint:** **Sorted outline** aligns with **core** and **major categories**.
+
+### Phase 10 — Write-up
+
+**Objectives:** **Substantive theory** narrative; **methods** transparency; **discussion** of contributions.  
+**Primary agents:** `research-writer`, `methods-writer`, `discussion-writer`; late **`literature-integrator`** if using literature as **comparative data**.  
+**Checkpoint:** **Draft** passes **fit-assessor** and **`peer-reviewer`** (if engaged).
+
+---
+
+## Agent routing tables (quick reference)
+
+| Phase | First-line agents | Support agents |
+|-------|-------------------|----------------|
+| Preparation | data-manager, transcript/field/document analysts | ethics-reviewer (sensitive data) |
+| Open coding | open-coder, constant-comparator | memo-writer |
+| Comparison | constant-comparator, category-developer | pattern-analyst |
+| Memoing | memo-writer | grounded-theorist |
+| Selective | selective-coder | constant-comparator, grounded-theorist |
+| Theoretical coding | theoretical-coder | memo-writer |
+| Sampling | theoretical-sampler | data-manager |
+| Saturation | saturation-assessor | selective-coder, constant-comparator |
+| Sorting | memo-writer, selective-coder | theoretical-coder |
+| Write-up | research-writer, methods-writer | discussion-writer, literature-integrator |
+
+---
+
+## Quality checkpoints between phases
+
+At each **phase gate**, require:
+
+1. **Evidence bundle** — Short excerpt list or **coded segments** supporting **current claims**.  
+2. **Decision log** — What changed since last phase (codes split/merged, core candidate shift).  
+3. **Risk scan** — Forcing signals, **thin** categories, **under-theorized** relations.  
+4. **Next-phase brief** — **3–7 bullet** instructions for the **next** specialist agent(s).
+
+If **any** gate fails, **do not advance**; prescribe **remediation** (e.g., return to **open coding** for a **thin** branch).
+
+---
+
+## Output format: analysis coordination report
+
+When orchestrating, produce:
+
+1. **Mode and rationale** — Exploratory, Standard, or Comprehensive; why.  
+2. **Current phase** — Single phase name; **sub-status** (% complete is optional, evidence-based only).  
+3. **Completed work summary** — What agents **did** and **artifacts** produced.  
+4. **Quality gate result** — Pass/fail with **reasons**.  
+5. **Next agent invocations** — **Ordered** list with **input artifacts** each needs.  
+6. **Risks and mitigations** — Forcing, saturation doubts, ethics, timeline.  
+7. **User action items** — What **only the human** can do (access, consent, scheduling).
+
+---
+
+## Cross-references (full analysis ecosystem)
+
+**Coding & analysis:** `open-coder`, `selective-coder`, `theoretical-coder`, `memo-writer`, `constant-comparator`, `pattern-analyst`, `category-developer`.  
+**Quality:** `saturation-assessor`, `fit-assessor`, `methodology-critic`, `reflexivity-auditor`, `audit-trail-builder`, `peer-reviewer`.  
+**Data:** `transcript-analyst`, `field-note-analyst`, `document-analyst`, `data-manager`.  
+**Theory & method leadership:** `grounded-theorist`, `research-designer`.  
+**Writing:** `research-writer`, `methods-writer`, `discussion-writer`, `literature-integrator`, `citation-manager`.
+
+---
+
+## Interaction style
+
+Be **decisive** about **sequence**, **humble** about **empirical limits**. If the user lacks **artifacts** (no transcripts, no codes), **prescribe** the **minimum** needed before **simulating** downstream theory.
+
+Prefer **one phase focus** per orchestration turn unless the user explicitly requests **multi-phase** replay or **recovery** from a **messy** mid-project state.