convoke-agents 2.0.0

package/_bmad/bme/_vortex/workflows/pattern-mapping/steps/step-01-setup.md

@@ -0,0 +1,102 @@
+---
+step: 1
+workflow: pattern-mapping
+title: Setup & Input Validation
+---
+
+# Step 1: Setup & Input Validation
+
+Before we map any patterns, we need to know exactly what research we're working with and assess whether it's ready for cross-source analysis.
+
+## Why This Matters
+
+One data point is an anecdote, three are a pattern. Pattern mapping only works when you have multiple research streams to compare. This step ensures we have enough material to identify meaningful cross-artifact patterns — and that we understand what each artifact brings to the table.
+
+## What Pattern Mapping Does
+
+Pattern mapping identifies cross-cutting themes and convergent patterns across multiple research artifacts. It's the **reconnaissance before convergence** — you're surveying what's there before committing to a full problem definition.
+
+**How it differs from Research Convergence:**
+- **Pattern mapping** identifies patterns (reconnaissance) — "What are the artifacts telling us?"
+- **Research convergence** produces a full HC2 problem definition (commitment) — "What is the problem?"
+
+Pattern mapping is lighter-weight. It helps you see whether your research converges before investing in full JTBD framing and Pains & Gains analysis.
+
+### When Pattern Mapping Helps
+
+- **You have 2+ artifacts but haven't synthesized them yet** — you want to see what patterns emerge before committing to full synthesis
+- **You suspect convergence but want to validate** — map patterns first to confirm your research streams agree
+- **You're identifying gaps before committing** — discover what's missing before running research-convergence
+
+Once patterns are clear, run Mila's **research-convergence** workflow as the next step to produce the full problem definition.
+
+## Your Task
+
+### 1. What Artifacts Do You Have?
+
+Mila accepts two categories of input for pattern mapping:
+
+**Primary: HC1 Empathy Artifacts from Isla**
+- Empathy maps (from Isla's `empathy-map` workflow)
+- Interview syntheses (from Isla's `user-interview` workflow)
+- Observation reports (from Isla's `user-discovery` workflow)
+
+**Also accepted: Any well-formed research input**
+- Informal research notes, external reports, prior pattern analyses
+- Non-Vortex research from other tools or methodologies
+- Minimum **2 artifacts recommended** for meaningful cross-source pattern identification
+
+### 2. List Your Input Artifacts
+
+Please provide the file paths or describe the artifacts you want to analyze. For example:
+- `_bmad-output/vortex-artifacts/hc1-empathy-map-busy-parents-2026-02-20.md`
+- `_bmad-output/vortex-artifacts/hc1-interview-synthesis-2026-02-21.md`
+- Or: "I have interview notes and survey results from an external tool"
+
+### 3. Input Assessment
+
+I'll assess each artifact using the appropriate mode:
+
+**For HC1-conforming artifacts — schema validation:**
+
+*Frontmatter check:*
+- `contract: HC1`
+- `type: artifact`
+- `source_agent` (who produced it)
+- `source_workflow` (which workflow)
+- `target_agents` includes `mila`
+- `input_artifacts` (upstream references)
+- `created` (date)
+
+*Body section check:*
+- Executive Summary
+- Research Context
+- Synthesized Insights
+- Key Themes
+- Pain Points
+- Desired Gains
+- Recommendations
+
+**For non-HC1 research — structural clarity assessment:**
+
+I'll check for minimal structure without rejecting:
+- **Source identification** — author, date, method if known
+- **Scope/context** — who was researched, what was studied
+- **Findings** — key discoveries
+- **Evidence summary** — how strong is the data
+
+**If artifacts are non-conforming:** That's okay — we don't reject research. I'll note which sections are present and which are missing, then ask clarification questions to fill gaps. The data shows that even partial research reveals patterns when we're explicit about what evidence we have and what we're working without.
+
+> For the full HC1 schema reference, see `{project-root}/_bmad/bme/_vortex/contracts/hc1-empathy-artifacts.md`
+
+---
+
+## Your Turn
+
+Please list your input artifacts — file paths, descriptions, or both. I'll assess them and we'll proceed to loading and analysis.
+
+## Next Step
+
+When your artifacts are listed and assessed, I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/pattern-mapping/steps/step-02-context.md

package/_bmad/bme/_vortex/workflows/pattern-mapping/steps/step-02-context.md

@@ -0,0 +1,81 @@
+---
+step: 2
+workflow: pattern-mapping
+title: Context Loading & Analysis
+---
+
+# Step 2: Context Loading & Analysis
+
+Now that we know what artifacts we're working with, let's load them all and understand what each one individually reveals before we start looking for cross-artifact patterns.
+
+## Why This Matters
+
+Before you can see patterns across artifacts, you need to understand what each artifact says on its own. The data shows that rushing to cross-artifact comparison without per-artifact analysis leads to shallow pattern identification. Take the time to understand each piece before looking at the whole picture.
+
+## Your Task
+
+### 1. Load All Input Artifacts
+
+Read through each of your input artifacts completely. Don't skim — the details matter for pattern identification.
+
+For each artifact, note:
+- **Who** was researched (user segments, personas, interview subjects)
+- **What** was discovered (key findings, insights, themes)
+- **How strong** the evidence is (direct quotes, observed behavior, survey data, or assumptions)
+- **Themes identified** within this single artifact
+- **Contradictions** with other artifacts (if you notice any while reading)
+
+### 2. Per-Artifact Analysis
+
+For each artifact, build a brief analysis note. Here's the structure:
+
+**Artifact: [Name/Path]**
+- **Source:** [Who produced it, when, using what method]
+- **Scope:** [Who was researched, what was studied]
+- **Top 3 Findings:** [The strongest insights from this artifact]
+- **Evidence Strength:** Strong (direct quotes/observed behavior) / Moderate (survey data/indirect evidence) / Weak (assumptions/anecdotal)
+- **Initial Themes:** [Themes this artifact surfaces on its own]
+
+### 3. Establish the Artifact Landscape
+
+Once you've analyzed each artifact individually, summarize the landscape:
+- **Strongest individual artifact:** Which artifact has the most robust evidence?
+- **Broadest coverage:** Which artifact covers the most ground?
+- **Most surprising content:** What did you find that you didn't expect?
+- **Obvious overlaps:** Even before formal pattern identification, what topics appear in multiple artifacts?
+
+---
+
+## Example Per-Artifact Analysis
+
+**From 2 artifacts** (empathy map + interview synthesis):
+
+**Artifact 1: HC1 Empathy Map — Busy Parents Segment**
+- **Source:** Isla's empathy-map workflow, 2026-02-20
+- **Scope:** 12 busy parents, ages 28-42, dual-income households
+- **Top 3 Findings:** (1) Morning routine is the highest-stress period; (2) Parents build workarounds for meal planning; (3) "I just want someone to tell me it's okay to not do everything"
+- **Evidence Strength:** Strong — direct quotes from empathy mapping sessions
+- **Initial Themes:** Time pressure, decision fatigue, guilt
+
+**Artifact 2: HC1 Interview Synthesis — Working Parents**
+- **Source:** Isla's user-interview workflow, 2026-02-21
+- **Scope:** 8 working parents, semi-structured interviews, 45 min each
+- **Top 3 Findings:** (1) 6 of 8 participants mention morning routine as pain point; (2) Existing tools feel designed for "people with more time"; (3) Partner coordination is a hidden complexity layer
+- **Evidence Strength:** Strong — direct quotes, behavioral observations
+- **Initial Themes:** Time pressure, tool inadequacy, coordination overhead
+
+**Artifact Landscape:**
+- **Strongest evidence:** Artifact 2 (longer interviews, richer quotes)
+- **Broadest coverage:** Artifact 1 (more participants, wider demographic)
+- **Most surprising:** Partner coordination theme — appeared strongly in interviews but was absent from empathy map
+- **Obvious overlaps:** Time pressure and morning routine pain appear in both
+
+## Your Turn
+
+Walk me through your per-artifact analysis. What does each artifact independently reveal? Three patterns converge on this insight — but first we need to understand the parts before we see the whole.
+
+## Next Step
+
+When we've established the per-artifact landscape, I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/pattern-mapping/steps/step-03-pattern-identification.md

package/_bmad/bme/_vortex/workflows/pattern-mapping/steps/step-03-pattern-identification.md

@@ -0,0 +1,88 @@
+---
+step: 3
+workflow: pattern-mapping
+title: Pattern Identification
+---
+
+# Step 3: Pattern Identification
+
+Now that we understand what each artifact individually reveals, it's time to look across all artifacts together and identify the patterns that no single artifact shows alone.
+
+## Why This Matters
+
+Cross-source triangulation is where the real insights emerge. One data point is an anecdote, three are a pattern. When the same theme, behavior, or pain appears across multiple independent research streams, that's a signal worth paying attention to. This step systematically maps those cross-artifact patterns.
+
+## Your Task
+
+### 1. Cross-Artifact Pattern Identification
+
+Look across all your artifacts and identify:
+
+**Recurring Themes** — What topics or concerns appear in multiple artifacts?
+- Example: "Both the empathy map and interview synthesis mention time pressure as a recurring theme"
+
+**Convergent Behaviors** — What user behaviors show up repeatedly?
+- Example: "Three separate artifacts reference users creating workarounds for the same limitation"
+
+**Contradictions** — Where do artifacts disagree or present conflicting evidence?
+- Example: "Empathy map suggests users want more features, but interviews reveal they feel overwhelmed by current complexity"
+
+**Evidence Gaps** — What questions remain unanswered across all artifacts?
+- Example: "We have strong evidence about pain points but almost no data on what users would consider a successful outcome"
+
+### 2. Pattern Strength Assessment
+
+For each pattern you identify, assess its strength:
+
+| Strength | Criteria | Implication |
+|----------|----------|-------------|
+| **Universal** | Appears in all artifacts with consistent evidence | Strong signal — likely a real pattern |
+| **Multi-source** | Appears in 2+ artifacts | Pattern emerging — worth investigating |
+| **Single** | Appears in only 1 artifact | Anecdote — needs more evidence before acting on it |
+
+### 3. Build Your Pattern Inventory
+
+Capture each pattern in this format:
+
+| Pattern | Source Artifacts | Strength | Type | Evidence Count |
+|---------|-----------------|----------|------|---------------|
+| *Name the pattern* | *Which artifacts* | *Universal / Multi-source / Single* | *Theme / Behavior / Contradiction / Gap* | *Number of data points* |
+
+**Pattern types:**
+- **Theme** — A recurring topic or concern (e.g., "time pressure")
+- **Behavior** — A repeated user action or coping mechanism (e.g., "building personal workarounds")
+- **Contradiction** — Conflicting evidence between artifacts (e.g., "want more features vs. overwhelmed by complexity")
+- **Gap** — Missing evidence or unanswered questions (e.g., "no data on success criteria")
+
+---
+
+## Example Pattern Inventory
+
+**From 3 artifacts** (empathy map, interview synthesis, observation report):
+
+| Pattern | Source Artifacts | Strength | Type | Evidence Count |
+|---------|-----------------|----------|------|---------------|
+| Morning routine is highest-stress period | Empathy map, Interviews, Observations | Universal | Theme | 18 data points across all 3 |
+| Users build personal workarounds (spreadsheets, sticky notes) | Interviews, Observations | Multi-source | Behavior | 8 instances across 2 artifacts |
+| Want more features vs. overwhelmed by complexity | Empathy map vs. Interviews | Multi-source | Contradiction | 5 + 6 data points in opposition |
+| No data on what "success" looks like to users | All 3 (by absence) | Universal | Gap | 0 — completely absent |
+
+Here's what the research is telling us: the first pattern (morning stress) has the strongest evidence — 18 data points across all three artifacts. That's not an anecdote; that's a clear signal. The contradiction (features vs. complexity) is also critical — it may reveal a deeper pattern about users who want capability without cognitive load.
+
+## Your Turn
+
+Build your pattern inventory. What patterns emerge when you lay your artifacts side by side? Across these artifacts, the evidence clusters around — share what you're finding.
+
+---
+
+**[a]** Advanced Elicitation — Deep dive into pattern identification with guided questioning
+**[p]** Party Mode — Bring in other Vortex agents for collaborative discussion
+**[c]** Continue — Move to theme clustering
+
+---
+
+## Next Step
+
+When your pattern inventory is complete, I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/pattern-mapping/steps/step-04-theme-clustering.md

package/_bmad/bme/_vortex/workflows/pattern-mapping/steps/step-04-theme-clustering.md

@@ -0,0 +1,100 @@
+---
+step: 4
+workflow: pattern-mapping
+title: Theme Clustering
+---
+
+# Step 4: Theme Clustering
+
+With your patterns identified, it's time to group them into coherent themes. Individual patterns are data points — theme clusters reveal the bigger story.
+
+## Why This Matters
+
+Patterns on their own are fragments. Theme clustering organizes those fragments into a coherent picture. The data shows that teams who cluster patterns before defining problems produce stronger, more focused problem statements. This step transforms your pattern inventory into a thematic map of what the research collectively reveals.
+
+## Your Task
+
+### 1. Group Related Patterns into Themes
+
+Look at your pattern inventory and ask: which patterns belong together? Which ones tell different parts of the same story?
+
+**Clustering criteria:**
+- Patterns that share the same root cause
+- Patterns that affect the same user need or job
+- Patterns that would be addressed by the same type of solution
+- Contradictions that illuminate different facets of the same issue
+
+### 2. Assess Theme Confidence
+
+For each theme cluster, assign a confidence level:
+
+| Confidence | Criteria |
+|------------|----------|
+| **Strong** | 3+ artifacts converge with consistent evidence — this theme is well-established |
+| **Moderate** | 2 artifacts converge, or 1 artifact with multiple independent data points |
+| **Emerging** | Single-source evidence, or contradictions exist within the theme |
+
+### 3. Build Your Theme Summary
+
+For each theme cluster, capture:
+
+**Theme: [Name]**
+- **Contributing Patterns:** [List the patterns from Step 3 that belong to this theme]
+- **Evidence Count:** [Total data points across all contributing patterns]
+- **Confidence:** Strong / Moderate / Emerging
+- **Key Insight:** [One sentence capturing what this theme tells us]
+- **Implications:** [What does this mean for problem definition?]
+
+### 4. Cross-Theme Synthesis
+
+Once your themes are clustered, step back and ask:
+
+- **Do the themes collectively point toward a problem definition?** If strong themes converge on a clear problem, you may be ready for research-convergence or even hypothesis-engineering.
+- **Do themes contradict each other?** Conflicting themes may signal that you're looking at multiple user segments or that more research is needed.
+- **Are there theme gaps?** Missing themes (e.g., strong evidence on pains but no evidence on desired outcomes) suggest where further research would be most valuable.
+- **Which themes are actionable?** Strong themes with clear implications can drive next steps. Emerging themes may need more evidence.
+
+---
+
+## Example Theme Clusters
+
+**From the pattern inventory in Step 3:**
+
+**Theme 1: Time as the Core Constraint**
+- **Contributing Patterns:** Morning routine stress (Universal), Personal workarounds (Multi-source)
+- **Evidence Count:** 26 data points (18 + 8)
+- **Confidence:** Strong
+- **Key Insight:** Users consistently experience time as their primary constraint — they build workarounds because existing tools don't respect their time limitations.
+- **Implications:** Any solution must reduce time-to-value. Features that add steps will fail regardless of utility.
+
+**Theme 2: Capability vs. Complexity Tension**
+- **Contributing Patterns:** Want more features vs. overwhelmed by complexity (Contradiction)
+- **Evidence Count:** 11 data points (5 + 6)
+- **Confidence:** Moderate
+- **Key Insight:** Users want more capability but experience current complexity as a burden. This isn't contradictory — they want capability without cognitive load.
+- **Implications:** The problem may be about information architecture, not feature set. Suggests progressive disclosure as a design principle.
+
+**Theme 3: Undefined Success Criteria**
+- **Contributing Patterns:** No data on what "success" looks like (Gap)
+- **Evidence Count:** 0 (gap identified by absence)
+- **Confidence:** Emerging
+- **Key Insight:** Across all research, we never asked users what a good outcome looks like — we only documented their pains.
+- **Implications:** Before defining the problem, we may need Isla to investigate what success means to users. Without this, any problem definition risks being pain-reactive rather than outcome-oriented.
+
+## Your Turn
+
+Group your patterns into theme clusters, assess confidence levels, and synthesize across themes. The patterns are starting to emerge — share what picture they paint.
+
+---
+
+**[a]** Advanced Elicitation — Deep dive into theme clustering with guided questioning
+**[p]** Party Mode — Bring in other Vortex agents for collaborative discussion
+**[c]** Continue — Move to synthesis and routing
+
+---
+
+## Next Step
+
+When your theme clusters are complete, I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/pattern-mapping/steps/step-05-synthesize.md

package/_bmad/bme/_vortex/workflows/pattern-mapping/steps/step-05-synthesize.md

@@ -0,0 +1,135 @@
+---
+step: 5
+workflow: pattern-mapping
+title: Synthesize & Route
+---
+
+# Step 5: Synthesize & Route
+
+Time to bring everything together. We've assessed artifacts, analyzed individual contributions, identified cross-artifact patterns, and clustered them into themes. Now we produce the pattern analysis output — a document that captures what the research collectively reveals.
+
+## Why This Matters
+
+This is reconnaissance — we're documenting what's there before committing to a problem definition. Here's what the research is telling us: across all your artifacts, the evidence clusters around certain themes. The pattern analysis output gives you (and your team) a clear picture of where your research converges, where it contradicts, and where the gaps are. One data point is an anecdote, three are a pattern — and this output captures which patterns have earned that distinction. It can feed into research-convergence for full synthesis, or stand alone as a valuable reference.
+
+## Your Task
+
+### 1. Pattern Analysis Summary
+
+Summarize the key findings from your pattern mapping:
+
+- **How many patterns did you identify?** (from Step 3)
+- **How many theme clusters emerged?** (from Step 4)
+- **What is the strongest signal?** — the theme with the highest confidence and most evidence
+- **What is the most significant contradiction?** — where does the evidence pull in different directions
+- **What is the biggest gap?** — what would change your conclusions if you had this data
+
+### 2. Generate the Pattern Analysis Artifact
+
+I'll produce the pattern analysis output with this structure:
+
+```yaml
+---
+source_agent: mila
+source_workflow: pattern-mapping
+target_agents: [mila, liam, isla]
+input_artifacts:
+  - path: "_bmad-output/vortex-artifacts/{your-artifact-1}"
+    contract: HC1
+  - path: "{path-to-your-artifact-2}"
+    contract: null
+created: YYYY-MM-DD
+---
+```
+
+**Note:** This is NOT an HC2 problem definition. There is no `contract` field — this is a working document for internal Vortex use, not a handoff contract. It can feed into research-convergence for full problem definition or stand alone as a reference.
+
+**Required Body Sections (6):**
+
+1. **Pattern Summary** — Strongest cross-artifact patterns with evidence sources
+   - List each pattern with: name, strength (Universal/Multi-source/Single), evidence count, and which artifacts support it
+
+2. **Theme Clusters** — Grouped patterns with confidence levels
+   - For each theme: name, contributing patterns, confidence (Strong/Moderate/Emerging), key insight
+
+3. **Convergence Assessment** — How strongly patterns point toward a unified picture
+   - Do patterns converge on a clear direction? How many themes are Strong vs. Emerging?
+   - Overall convergence: High (most themes Strong) / Moderate (mix of Strong and Moderate) / Low (mostly Emerging or contradictory)
+
+4. **Contradictions** — Conflicting evidence and possible explanations
+   - List each contradiction with: what conflicts, which artifacts, possible reconciliation
+
+5. **Evidence Gaps** — What's missing
+   - List gaps that would change conclusions if filled
+   - Note which gaps block confident routing vs. which are nice-to-have
+
+6. **Recommendations** — Suggested next steps based on pattern strength
+   - Which Compass route does the evidence support? Why?
+   - What would strengthen the analysis?
+
+**Save to:** `{output_folder}/vortex-artifacts/pattern-analysis-{date}.md`
+
+I'll create this file with all sections once you confirm the content is ready.
+
+### 3. Validation Questions
+
+Before we finalize, let's validate:
+
+**Evidence Check:**
+- [ ] Is every pattern backed by at least one artifact reference?
+- [ ] Are confidence levels honest given the evidence count?
+- [ ] Did we document contradictions and gaps rather than hiding them?
+
+**Completeness Check:**
+- [ ] Does the pattern summary capture all significant cross-artifact patterns?
+- [ ] Do theme clusters account for all identified patterns?
+- [ ] Is the convergence assessment consistent with theme confidence levels?
+
+---
+
+## Your Turn
+
+Let's build the Pattern Analysis Summary and confirm the content for each section. Once you're ready, I'll generate the artifact.
+
+---
+
+**[a]** Advanced Elicitation — Deep dive into synthesis with guided questioning
+**[p]** Party Mode — Bring in other Vortex agents for collaborative discussion
+**[c]** Continue — Generate the pattern analysis artifact and proceed to routing
+
+---
+
+## Vortex Compass
+
+Based on what your pattern analysis reveals, here are your evidence-driven options:
+
+| If you learned... | Consider next... | Agent | Why |
+|---|---|---|---|
+| Patterns identified and ready for full synthesis | research-convergence | Mila 🔬 | Patterns provide strong foundation for full JTBD + Pains & Gains problem definition (HC2) |
+| Patterns already point to a clear, well-evidenced problem | hypothesis-engineering | Liam 💡 | Sufficient evidence to skip synthesis and go directly to hypothesis generation |
+| Patterns reveal significant knowledge gaps | user-discovery | Isla 🔍 | More discovery needed to fill gaps before synthesis is meaningful |
+
+**Self-routing note:** Route 1 sends you back to Mila's own research-convergence workflow. This is the expected successful completion path — pattern mapping is reconnaissance, research-convergence is commitment. This is not a failure state.
+
+> **Note:** These are evidence-based recommendations. You can navigate to any Vortex agent
+> at any time based on your judgment.
+
+**Or run Max's [VN] Vortex Navigation** for a full gap analysis across all streams.
+
+### ⚠️ Insufficient Evidence for Routing
+
+If the evidence gathered so far doesn't clearly point to a single next step:
+
+| To route to... | You need... |
+|----------------|-------------|
+| Mila 🔬 research-convergence | Identified patterns with at least Moderate confidence themes that can ground JTBD framing |
+| Liam 💡 hypothesis-engineering | Strong convergence across themes pointing to a clear, well-evidenced problem |
+| Isla 🔍 user-discovery | Identified gaps or contradictions that block confident problem definition |
+
+**Workflow-specific signals:**
+- Patterns don't converge on any clear theme → more artifacts needed, consider **Isla** for focused research
+- Contradictions dominate over convergence → **Isla** for focused investigation to resolve conflicting evidence
+- Strong patterns but unclear which job they serve → proceed to **research-convergence** for JTBD framing
+- Fewer than 2 artifacts with converging insights → patterns emerge from cross-artifact convergence. Consider collecting more artifacts or using **Isla's** workflows to deepen existing research before proceeding.
+
+**Recommended:** If routing is unclear, revisit your theme clusters from Step 4 or run **Max's [VN] Vortex Navigation** for a full gap analysis.

package/_bmad/bme/_vortex/workflows/pattern-mapping/workflow.md

@@ -0,0 +1,58 @@
+---
+workflow: pattern-mapping
+type: step-file
+description: Map cross-source patterns across research artifacts to identify convergent themes before full synthesis
+author: Mila (research-convergence-specialist)
+version: 1.6.0
+---
+
+# Pattern Mapping Workflow
+
+This workflow guides you through mapping cross-source patterns across multiple research artifacts to identify convergent themes, contradictions, and evidence gaps — before committing to a full problem definition.
+
+## What is Pattern Mapping?
+
+Pattern mapping is the reconnaissance before convergence. Where research-convergence produces a full HC2 problem definition (commitment), pattern mapping identifies cross-cutting patterns and themes across your research artifacts (reconnaissance). It answers: "What are the artifacts collectively telling us?" without yet committing to a formal problem statement.
+
+Here's what the research is telling us: you don't always need to jump straight to problem definition. Sometimes the most valuable step is laying your research side by side and seeing what patterns emerge. That's what this workflow does.
+
+**When Pattern Mapping Helps:**
+- You have 2+ artifacts but haven't synthesized them yet — you want to see what's there before committing
+- You suspect convergence but want to map patterns before investing in a full problem definition
+- You want to validate that research streams agree before running research-convergence
+- You're identifying gaps that need more Isla research before synthesis is worthwhile
+
+Once patterns are clear, run **research-convergence** to produce the full HC2 problem definition.
+
+## Workflow Structure
+
+**Step-file architecture:**
+- Just-in-time loading (each step loads only when needed)
+- Sequential enforcement (must complete step N before step N+1)
+- State tracking in frontmatter (progress preserved)
+
+## Steps Overview
+
+1. **Setup & Input Validation** — Validate your input artifacts (HC1 empathy artifacts or any well-formed research)
+2. **Context Loading & Analysis** — Load artifacts and analyze what each one individually reveals
+3. **Pattern Identification** — Identify cross-artifact patterns: recurring themes, convergent behaviors, contradictions, evidence gaps
+4. **Theme Clustering** — Group related patterns into coherent themes with confidence levels
+5. **Synthesize & Route** — Produce a pattern analysis output and route via Compass
+
+## Output
+
+**Artifact:** Pattern analysis markdown file in `{output_folder}/vortex-artifacts/pattern-analysis-{date}.md`
+
+**Template:** None
+
+**Schema:** This is a working document — NOT an HC2 problem definition. No handoff contract applies. The output can feed into research-convergence for full synthesis or stand alone as a pattern analysis reference.
+
+**Consumers:** Mila (research-convergence), Liam (hypothesis-engineering), Isla (user-discovery) — depending on Compass routing.
+
+---
+
+## INITIALIZATION
+
+Load config from {project-root}/_bmad/bme/_vortex/config.yaml
+
+Load step: {project-root}/_bmad/bme/_vortex/workflows/pattern-mapping/steps/step-01-setup.md