convoke-agents 2.0.0

package/_bmad/bme/_vortex/workflows/pivot-patch-persevere/workflow.md

@@ -0,0 +1,51 @@
+---
+workflow: pivot-patch-persevere
+type: step-file
+description: Make strategic pivot, patch, or persevere decisions based on experiment evidence
+author: Max (learning-decision-expert)
+version: 1.5.0
+---
+
+# Pivot / Patch / Persevere Decision Workflow
+
+This workflow guides you through making a rigorous strategic decision after experiments: should you change direction (Pivot), adjust your approach (Patch), or stay the course (Persevere)?
+
+## What is Pivot / Patch / Persevere?
+
+Based on Eric Ries' Lean Startup methodology, this is a structured decision framework for what to do after experiments. Instead of relying on gut feelings or sunk-cost reasoning, you systematically evaluate evidence and make a defensible choice:
+
+- **Pivot** - Fundamentally change direction. Your core hypothesis is wrong, and incremental adjustments won't fix it. You need a different approach, market, value proposition, or business model.
+- **Patch** - Adjust your approach while keeping the core direction. The hypothesis is partially right but your execution, targeting, or specific solution needs modification.
+- **Persevere** - Stay the course. Evidence supports your direction, and you should double down with more resources, broader experiments, or next-stage development.
+
+**Why these three, not just "pivot or persevere"?** Most real decisions aren't binary. The most common outcome is "we're directionally right but need to adjust." Patch captures this middle ground and prevents both premature pivots (throwing away good ideas) and false perseverance (refusing to adapt).
+
+## 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. **Evidence Review** - Gather all relevant learning cards, experiment results, and evidence
+2. **Hypothesis Assessment** - Assess original hypotheses: confirmed, partially confirmed, or invalidated?
+3. **Option Analysis** - Analyze all three options with pros, cons, and risks
+4. **Stakeholder Input** - Capture team perspectives, concerns, and preferences
+5. **Decision** - Make and document the decision with clear rationale
+6. **Action Plan** - Create a concrete action plan for the chosen direction
+
+## Output
+
+**Artifact:** Decision record in `{output_folder}/ppp-decision-{topic}-{date}.md`
+
+**Template:** Uses [pivot-patch-persevere.template.md](pivot-patch-persevere.template.md)
+
+---
+
+## INITIALIZATION
+
+Load config from {project-root}/_bmad/bme/_vortex/config.yaml
+
+Load step: {project-root}/_bmad/bme/_vortex/workflows/pivot-patch-persevere/steps/step-01-evidence-review.md

package/_bmad/bme/_vortex/workflows/pivot-resynthesis/steps/step-01-setup.md

@@ -0,0 +1,97 @@
+---
+step: 1
+workflow: pivot-resynthesis
+title: Setup & Input Validation
+---
+
+# Step 1: Setup & Input Validation
+
+You're here because Max analyzed the evidence and made a decision: **pivot**. The solution direction failed, but the problem definition is sound. The experiment told us something important — and now we use that knowledge to sharpen, not restart.
+
+## Why This Matters
+
+Pivot resynthesis is fundamentally different from fresh convergence. You already have a validated problem space. What you're doing now is integrating new evidence — from experiments that didn't go as expected — into a revised problem definition. We're not starting over. We're iterating within the known space.
+
+## Your Task
+
+### 1. Understand the Pivot Context
+
+Max decided "pivot" based on the evidence. This means:
+- **The problem is correct** — users do experience the pain your research identified
+- **The solution direction failed** — the hypothesis about how to solve it was wrong
+- **New evidence exists** — the failed experiment revealed what doesn't work and why
+
+### 2. List Your Input Artifacts
+
+You'll need two types of input:
+
+**Original Research (HC1 Empathy Artifacts):**
+Your original Isla artifacts — empathy maps, interview syntheses, observation reports. These provide the research foundation that remains valid.
+
+**Experiment Evidence (HC4 or informal):**
+Results from the failed experiment — what was tested, what happened, what was learned. This can be a formal HC4 artifact from Wade or informal experiment notes.
+
+**Prior Problem Definition (optional):**
+If you have the original HC2 problem definition from before the experiment, it provides the JTBD baseline for re-framing.
+
+### 3. Validate Your Artifacts
+
+**HC1 Empathy Artifacts — Frontmatter Check:**
+- `contract: HC1`
+- `type: artifact`
+- `source_agent` (e.g., `isla`)
+- `source_workflow` (e.g., `empathy-map`, `user-interview`, `user-discovery`)
+- `target_agents` (should include `mila`)
+- `input_artifacts`
+- `created` (YYYY-MM-DD)
+
+**HC1 Body Sections:**
+- Executive Summary
+- Research Context
+- Synthesized Insights
+- Key Themes
+- Pain Points
+- Desired Gains
+- Recommendations
+
+Reference: `{project-root}/_bmad/bme/_vortex/contracts/hc1-empathy-artifacts.md`
+
+**HC4 Experiment Context — Frontmatter Check:**
+- `contract: HC4`
+- `type: artifact`
+- `source_agent` (e.g., `wade`)
+- `source_workflow` (e.g., `lean-experiment`, `proof-of-concept`)
+- `target_agents`
+- `input_artifacts`
+- `created` (YYYY-MM-DD)
+
+**HC4 Body Sections:**
+- Experiment Summary
+- Hypothesis Tested
+- Experiment Method
+- Pre-Defined Success Criteria
+- Confirmed/Rejected Hypotheses
+- Strategic Context
+- Production Readiness (if experiment graduated to production)
+
+Reference: `{project-root}/_bmad/bme/_vortex/contracts/hc4-experiment-context.md`
+
+### 4. Non-Conforming Input
+
+If your artifacts don't perfectly match the HC1 or HC4 schemas, we don't reject them — the data shows that real-world research rarely arrives in perfect format. Instead, note which sections are present and which are missing. We'll work with what you have and guide you to fill gaps where needed.
+
+The critical information for pivot resynthesis is:
+- From HC1: What pains and gains did the original research identify?
+- From HC4: What was tested? What failed? What did we learn?
+
+---
+
+## Your Turn
+
+Walk me through your artifacts. What HC1 research do you have, and what experiment evidence are you bringing to the table? The data shows that the most effective pivots happen when we lay all the evidence side by side.
+
+## Next Step
+
+When we've validated your inputs, I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/pivot-resynthesis/steps/step-02-context.md

package/_bmad/bme/_vortex/workflows/pivot-resynthesis/steps/step-02-context.md

@@ -0,0 +1,86 @@
+---
+step: 2
+workflow: pivot-resynthesis
+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 build the pivot context. The data shows that the most important step in pivot resynthesis is understanding what the experiment actually revealed — and separating that from what the original research already told us.
+
+## Why This Matters
+
+You have two evidence streams: the original research and the experiment results. Laying them side by side reveals what still holds, what's been disproved, and what's newly understood. This is where we figure out what to revise and what to preserve.
+
+## Your Task
+
+### 1. Load All Input Artifacts
+
+Read through each of your input artifacts completely. For each artifact, note:
+- **What** was the original finding or hypothesis
+- **How strong** the evidence was (direct quotes, observed behavior, survey data, or assumptions)
+- **What happened** in the experiment related to this finding
+
+### 2. Establish Pivot Context
+
+Answer these four questions — they form the foundation for everything that follows:
+
+**(a) What was the original problem definition?**
+What JTBD, pains, and gains did the original HC2 define? If you don't have a prior HC2, summarize the original problem understanding from your HC1 artifacts.
+
+**(b) What hypothesis was tested?**
+What did the team believe would work? What riskiest assumption was the experiment designed to validate?
+
+**(c) What did the experiment reveal?**
+What failed and why? Was the hypothesis rejected outright, or partially confirmed? What unexpected results emerged?
+
+**(d) What unexpected findings emerged?**
+The data shows that failed experiments often reveal more than successful ones. What did you learn that nobody predicted?
+
+### 3. "What Still Holds?" Assessment
+
+This is the critical pivot analysis. For each element of your original problem definition, assess:
+
+| Element | Status | Evidence |
+|---------|--------|----------|
+| JTBD core (situation/motivation/outcome) | `Holds` / `Needs revision` / `Invalidated` | What evidence supports this assessment? |
+| Pain 1: [name] | `Retained` / `Revised` / `Removed` / `New` | Experiment confirmed/contradicted this? |
+| Pain 2: [name] | `Retained` / `Revised` / `Removed` / `New` | Experiment confirmed/contradicted this? |
+| Gain 1: [name] | `Retained` / `Revised` / `Reprioritized` / `Removed` / `New` | What did the experiment show about this gain? |
+| *Add rows as needed* | | |
+
+**New elements to add:**
+- Pains newly revealed by experiment failure
+- Gains suggested by unexpected experiment findings
+- Assumptions invalidated by experiment evidence
+
+### 4. Synthesis Inventory
+
+Based on your pivot context analysis, organize your thinking:
+
+- **Strongest surviving signal:** Which original finding is most strongly confirmed by the experiment evidence?
+- **Biggest revision needed:** What did the experiment most clearly disprove or change?
+- **Most surprising insight:** What unexpected learning emerged from the failed experiment?
+- **Revised problem shape:** If you had to describe the revised core problem in one sentence right now, what would it be?
+
+---
+
+## Example Pivot Context Analysis
+
+**From original research (HC1)** + **experiment results (HC4):**
+
+- **Original JTBD:** "When switching between tools during analysis, I want consolidated data views, so I can make decisions without context-switching."
+- **Experiment tested:** A consolidated dashboard reducing tool-switching by 50%
+- **What failed:** Users still switched tools despite the dashboard — the pain wasn't tool-switching, it was *not knowing which data to trust*
+- **What holds:** JTBD core holds (users want to make decisions without friction). Situation needs revision: not "switching between tools" but "encountering conflicting data."
+- **What changed:** Pain A (time spent switching tools) → Invalidated. New Pain: uncertainty about data accuracy. Gain priority shifted: "faster decisions" is less important than "confident decisions."
+
+## Your Turn
+
+Walk me through your pivot context analysis. The original research still holds — here's what we're revising based on what the experiment revealed. Three patterns converge on this insight — share what you're finding and I'll help you sharpen the revision.
+
+## Next Step
+
+When we've mapped the pivot landscape, I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/pivot-resynthesis/steps/step-03-jtbd-reframing.md

package/_bmad/bme/_vortex/workflows/pivot-resynthesis/steps/step-03-jtbd-reframing.md

@@ -0,0 +1,88 @@
+---
+step: 3
+workflow: pivot-resynthesis
+title: JTBD Re-Framing
+---
+
+# Step 3: JTBD Re-Framing
+
+Now we take your "What Still Holds?" assessment and apply it to the Jobs-to-be-Done framing. The JTBD core is sound. What changed is the evidence around it — and that may require revisions to the situation, motivation, or outcome.
+
+## Why This Matters
+
+The data shows that in a pivot scenario, the JTBD core typically remains stable. Max routed you here — not to Emma — because the problem itself is right. What failed is the solution direction, which means the way we framed the pains and gains around the job may need updating. But we start from the existing JTBD, not from blank.
+
+We're not starting over — we're sharpening.
+
+## Your Task
+
+### 1. Review the Original JTBD
+
+Start from your existing JTBD (from the original HC2 or your step-02 analysis):
+
+> When [original situation], I want to [original motivation], so I can [original expected outcome].
+
+### 2. Assess Each Component Against Experiment Evidence
+
+For each part of the JTBD, ask:
+
+**Situation — Does the triggering context still hold?**
+- Did the experiment confirm users are in this situation?
+- Or did it reveal the real triggering context is different?
+- Example: "When switching between tools" → revised to "When encountering conflicting data sources"
+
+**Motivation — Does the core want still hold?**
+- The motivation is usually the most stable part. Did the experiment challenge what users actually want?
+- Example: "I want consolidated data views" → typically stable, but might shift to "I want trustworthy data"
+
+**Expected Outcome — Does the desired end state still hold?**
+- Did the experiment reveal that users' real desired outcome is different from what was assumed?
+- Example: "so I can make decisions without context-switching" → revised to "so I can make decisions with confidence"
+
+### 3. Draft the Revised JTBD
+
+> When [revised situation], I want to [revised motivation], so I can [revised expected outcome].
+
+For each component you revised, note:
+- **What changed** and **why** (link to specific experiment evidence)
+- **What stayed the same** and **why** it's confirmed
+
+### 4. Assess Functional, Emotional, and Social Jobs
+
+**Functional Job** (required): What practical task is the user trying to accomplish?
+- Did the experiment change our understanding of the functional job?
+- The data shows that functional jobs rarely change in a pivot — the "what" is usually right, it's the "how" that failed.
+
+**Emotional Job** (if evidence exists): How does the user want to feel?
+- Failed experiments often reveal emotional jobs more clearly than successful ones.
+- Did the experiment surface emotional needs that the original research missed?
+
+**Social Job** (if evidence exists): How does the user want to be perceived?
+- Did experiment evidence surface social dynamics that affect the job?
+
+### 5. Evidence Grounding
+
+For every revision, ground it in specific evidence:
+- "We revised the situation because HC4 showed [specific finding]"
+- "The motivation holds because HC1 artifacts A and B both support [evidence]"
+- "The outcome shifted because the experiment revealed [unexpected result]"
+
+---
+
+## Your Turn
+
+Share your revised JTBD. Walk me through what changed, what held, and what evidence drove each decision. Across the original research and the pivot evidence, the revised problem should feel like a sharper version of the original — not a different problem entirely.
+
+---
+
+**[a]** Advanced Elicitation — Deep dive into JTBD re-framing with guided questioning
+**[p]** Party Mode — Bring in other Vortex agents for collaborative discussion
+**[c]** Continue — Move to Pains & Gains Revision
+
+---
+
+## Next Step
+
+When we've locked in the revised JTBD, I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/pivot-resynthesis/steps/step-04-pains-gains-revision.md

package/_bmad/bme/_vortex/workflows/pivot-resynthesis/steps/step-04-pains-gains-revision.md

@@ -0,0 +1,76 @@
+---
+step: 4
+workflow: pivot-resynthesis
+title: Pains & Gains Revision
+---
+
+# Step 4: Pains & Gains Revision
+
+With the revised JTBD in place, it's time to update the pains and gains. The experiment evidence changes the landscape — some original pains are confirmed, some are invalidated, and new ones have emerged. The data shows that the pains and gains table is where pivots have their biggest impact.
+
+## Why This Matters
+
+Pains and gains are the operational layer of your problem definition. They're what Liam will use to generate new hypotheses, and what Wade will use to design new experiments. Getting them right — grounded in both original research AND experiment evidence — is the highest-leverage activity in pivot resynthesis.
+
+## Your Task
+
+### 1. Revise Pains
+
+Work through each pain from your original problem definition using this framework:
+
+**Retain:** Pains validated by the experiment (evidence confirms users experience this)
+**Remove or deprioritize:** Pains invalidated by experiment evidence (the experiment showed this isn't actually a pain, or it's far less severe than assumed)
+**Add:** New pains revealed by the experiment failure (things we didn't know were painful until the experiment failed)
+**Update:** Pains whose frequency or intensity changed based on experiment data
+
+| Pain Description | Revision Status | Priority | Frequency | Intensity | Evidence Sources | Current Coping |
+|-----------------|-----------------|----------|-----------|-----------|-----------------|----------------|
+| *[Original pain — keep/revise/remove]* | `Retained` / `Revised` / `Removed` / `New` | High/Med/Low | How often | How severely | HC1 artifact + HC4 finding | How users cope |
+
+**For each retained pain:** Note which HC1 artifact originally identified it AND what HC4 evidence confirmed it.
+
+**For each removed pain:** Note which HC4 evidence invalidated it and why.
+
+**For each new pain:** Note which HC4 evidence revealed it and why the original research missed it.
+
+### 2. Revise Gains
+
+Apply the same framework to gains:
+
+**Retain:** Gains still relevant after the experiment
+**Reprioritize:** Gains whose importance shifted based on experiment results
+**Add:** New gains suggested by the pivot direction or experiment insights
+**Remove:** Gains that the experiment showed users don't actually want
+
+| Gain Description | Revision Status | Priority | Expected Impact | Evidence Sources |
+|-----------------|-----------------|----------|-----------------|-----------------|
+| *[Original gain — keep/revise/remove]* | `Retained` / `Revised` / `Reprioritized` / `Removed` / `New` | High/Med/Low | How this changes the user's situation | HC1 artifact + HC4 finding |
+
+### 3. Cross-Reference Against Revised JTBD
+
+Now verify alignment. For each revised pain and gain:
+- Does it connect logically to the revised JTBD from Step 3?
+- If a pain no longer applies to the job, remove it — even if it was originally High priority
+- If a gain doesn't relate to the revised motivation or outcome, reconsider its relevance
+
+Three patterns converge on this insight: the pains and gains that survive both original research AND experiment evidence are your strongest signals. These are the foundation for Liam's next round of hypotheses.
+
+---
+
+## Your Turn
+
+Walk me through your revised pains and gains tables. For each change, tell me: what changed, what evidence drove the change, and how it connects to the revised JTBD. Based on the experiment evidence, the revised picture should feel sharper and more grounded than the original.
+
+---
+
+**[a]** Advanced Elicitation — Deep dive into pains and gains revision with guided questioning
+**[p]** Party Mode — Bring in other Vortex agents for collaborative discussion
+**[c]** Continue — Move to Synthesis & Routing
+
+---
+
+## Next Step
+
+When we've finalized the revised pains and gains, I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/pivot-resynthesis/steps/step-05-synthesize.md

package/_bmad/bme/_vortex/workflows/pivot-resynthesis/steps/step-05-synthesize.md

@@ -0,0 +1,158 @@
+---
+step: 5
+workflow: pivot-resynthesis
+title: Synthesize & Route
+---
+
+# Step 5: Synthesize & Route
+
+Time to bring everything together. We've validated artifacts, established pivot context, re-framed the JTBD, and revised pains and gains. Now we converge it all into a revised HC2 Problem Definition artifact. Across the original research and the pivot evidence, the revised problem should feel like a sharper, evidence-strengthened version of the original.
+
+## Why This Matters
+
+This revised problem definition is what Liam will use to generate new hypotheses. It needs to be clear about what changed, what held, and why — so Liam doesn't repeat the failed experiment direction. The data shows that the most effective pivots are ones where the revised problem definition makes the old direction's failure obvious in hindsight.
+
+## Your Task
+
+### 1. Revised Converged Problem Statement
+
+Write a single, clear problem statement that synthesizes both the original research and the experiment evidence:
+
+**Problem Statement:** One concise paragraph defining the revised core problem. This should feel like a sharper version of the original — not a different problem entirely.
+
+**What Changed:** Explicitly note what shifted from the original problem definition and why.
+
+**Confidence:** `High` / `Medium` / `Low`
+- **High:** Original research confirmed by experiment evidence; revisions are evidence-backed
+- **Medium:** Evidence points in this direction but the pivot introduced new uncertainties
+- **Low:** Experiment contradicted significant original assumptions; the revised definition is best-effort
+
+Note: Confidence may increase (we eliminated wrong directions) or decrease (experiment revealed we misunderstood key pains). Assess honestly based on total evidence.
+
+**Scope:** What is explicitly in and out of scope for this revised problem.
+
+### 2. Evidence Summary
+
+Build the traceability record — how did we get here, and what's different?
+
+| Field | Your Answer |
+|-------|-------------|
+| **Artifacts Analyzed** | Count and list both HC1 artifacts AND HC4 experiment evidence used |
+| **Total Evidence Points** | Count discrete quotes, observations, and data points across all artifacts (original + experiment) |
+| **Convergence Assessment** | How strongly does the revised evidence converge on a single problem? Is it stronger or weaker than the original? |
+| **Contradictions** | Where did original research and experiment evidence conflict? How was this resolved? |
+| **Evidence Gaps** | What's still missing that would strengthen this revised problem definition? |
+
+**Provenance — What Changed and Why:**
+For each major revision to pains and gains, note:
+- Which HC4 experiment evidence triggered the revision
+- Which HC1 original source the item came from
+- Whether the item was retained, revised, removed, or newly added
+
+### 3. Assumptions
+
+Every problem definition embeds assumptions. After a pivot, some assumptions are validated, some invalidated, and some are new:
+
+| Assumption | Basis | Risk if Wrong | Validation Status |
+|------------|-------|---------------|-------------------|
+| *What we're assuming is true* | *Why it seems reasonable given original + experiment evidence* | *What happens if this is incorrect* | `Validated` / `Partially Validated` / `Assumed` |
+
+> **Pivot-resynthesis note:** For assumptions that were **invalidated by the experiment**, don't include them in this table — they've already been removed or revised in your pains/gains. This table captures what you're carrying forward. If you need to document what was invalidated, note it in the Evidence Summary's provenance section.
+
+Note which assumptions were **validated by the experiment** (strongest basis), which were **invalidated** (removed or revised), and which are **new assumptions** that emerged from the pivot evidence.
+
+### 4. Generate the HC2 Artifact
+
+I'll produce the revised HC2 Problem Definition artifact with this structure:
+
+```yaml
+---
+contract: HC2
+type: artifact
+source_agent: mila
+source_workflow: pivot-resynthesis
+target_agents: [liam]
+input_artifacts:
+  - path: "_bmad-output/vortex-artifacts/{your-hc1-artifact-1}"
+    contract: HC1
+  - path: "_bmad-output/vortex-artifacts/{your-hc4-experiment}"
+    contract: HC4
+created: YYYY-MM-DD
+---
+```
+
+**HC2 Required Body Sections:**
+1. **Converged Problem Statement** — Revised Problem Statement, What Changed, Confidence, Scope
+2. **Jobs-to-be-Done** — Revised JTBD (from Step 3) + Functional/Emotional/Social Jobs
+3. **Pains** — Revised pain table with revision status (from Step 4)
+4. **Gains** — Revised gain table with revision status (from Step 4)
+5. **Evidence Summary** — Dual-source traceability record (from above)
+6. **Assumptions** — Updated assumptions with validation status (from above)
+
+**Save to:** `{output_folder}/vortex-artifacts/hc2-problem-definition-{date}.md`
+
+I'll create this file with all the sections above once you confirm the content is ready.
+
+### 5. Validation Questions
+
+Before we finalize, let's validate:
+
+**Evidence Check:**
+- [ ] Is every revised pain and gain backed by at least one HC1 or HC4 artifact reference?
+- [ ] Can we trace the revised JTBD back to specific original research AND experiment evidence?
+- [ ] Did we document contradictions between original research and experiment evidence honestly?
+- [ ] Is it clear which items were retained, revised, removed, or newly added?
+
+**Convergence Check:**
+- [ ] Does the revised problem statement follow naturally from the revised JTBD + Pains + Gains?
+- [ ] Would someone reading only the revised problem statement understand what changed and why?
+- [ ] Is the confidence level honest given both original and experiment evidence?
+
+**Actionability Check:**
+- [ ] Can Liam use this to generate NEW hypotheses that avoid the failed experiment direction?
+- [ ] Are the assumptions specific enough to validate or invalidate?
+- [ ] Is it clear what the experiment disproved so Liam doesn't repeat the same direction?
+
+---
+
+## Your Turn
+
+Let's build the Revised Converged Problem Statement, Evidence Summary, and Assumptions table. Confirm when you're ready for me to generate the final HC2 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 HC2 artifact and proceed to routing
+
+---
+
+## Vortex Compass
+
+Based on what you just completed, here are your evidence-driven options:
+
+| If you learned... | Consider next... | Agent | Why |
+|---|---|---|---|
+| Revised problem definition is ready with strong evidence | hypothesis-engineering | Liam 💡 | Revised problem ready for new hypothesis generation (HC2) |
+| Assumptions from pivot need validation before new hypotheses | user-interview | Isla 🔍 | New assumptions need real-user validation before proceeding |
+
+> **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... |
+|----------------|-------------|
+| Liam 💡 | Revised problem definition with clear JTBD, revised pains/gains, and evidence that avoids the failed direction |
+| Isla 🔍 | Identified new assumptions from the pivot that are too risky to proceed without validation |
+
+**Workflow-specific signals:**
+- Revised problem statement is low-confidence → consider **Isla** for more evidence before proceeding
+- Experiment evidence contradicts all original pains → more discovery needed; consider **Isla** for focused research
+- JTBD itself is questioned (not just pains/gains) → the problem space may be wrong; consider **Emma** for re-scoping
+
+**Recommended:** Revisit earlier steps to strengthen your evidence, or run **Max's [VN] Vortex Navigation** for a full gap analysis.

package/_bmad/bme/_vortex/workflows/pivot-resynthesis/workflow.md

@@ -0,0 +1,52 @@
+---
+workflow: pivot-resynthesis
+type: step-file
+description: Re-synthesize a problem definition after failed experiments using original research plus pivot evidence
+author: Mila (research-convergence-specialist)
+version: 1.6.0
+---
+
+# Pivot Resynthesis Workflow
+
+This workflow guides you through re-synthesizing a problem definition after Max's pivot decision. You'll integrate your original Isla artifacts with new evidence from failed experiments to produce a revised, evidence-backed problem statement.
+
+## What is Pivot Resynthesis?
+
+The experiment told us something important: the solution direction was wrong, but the problem definition is sound. Pivot resynthesis is how we respond — not by starting over, but by sharpening what we already know.
+
+Here's what the research is telling us: your original findings still hold. What changed is the evidence landscape. Failed experiments reveal which pains and gains were real and which were assumptions. This workflow helps you revise your problem definition using both the original research and the new experiment evidence.
+
+The key distinction: research convergence creates a problem definition from scratch. Pivot resynthesis revises an existing one. We're not starting over — we're iterating within the known problem space.
+
+## 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 dual inputs: original HC1 empathy artifacts plus experiment evidence (HC4 or informal)
+2. **Context Loading & Analysis** - Load all artifacts, establish pivot context, assess "What Still Holds?"
+3. **JTBD Re-Framing** - Revise the existing JTBD based on experiment evidence (preserve what's valid, revise what's disproved)
+4. **Pains & Gains Revision** - Update pains and gains: retain validated, remove invalidated, add newly revealed
+5. **Synthesize & Route** - Converge into a revised HC2 problem definition and route via Compass
+
+## Output
+
+**Artifact:** HC2 Problem Definition markdown file in `{output_folder}/vortex-artifacts/hc2-problem-definition-{date}.md`
+
+**Template:** None (HC2 artifact is generated inline during Step 5)
+
+**Schema:** Conforms to HC2 contract (`_bmad/bme/_vortex/contracts/hc2-problem-definition.md`)
+
+**Consumer:** Liam (hypothesis-engineering) uses this to ground new hypothesis generation in a revised, evidence-strengthened problem.
+
+---
+
+## INITIALIZATION
+
+Load config from {project-root}/_bmad/bme/_vortex/config.yaml
+
+Load step: {project-root}/_bmad/bme/_vortex/workflows/pivot-resynthesis/steps/step-01-setup.md