refactor: rename cluster to behavior groups for clearer AI focus

Author: nullhack

.opencode/skills/scope/SKILL.md

@@ -63,7 +63,7 @@ Ask the stakeholder to describe the same situation from another actor's point of
 Three levels of active listening apply throughout every interview session:
 
 - **Level 1 — Per answer**: immediately paraphrase each answer before moving to the next question. "So if I understand correctly, you're saying that X happens when Y?" Catches misunderstanding in the moment.
-- **Level 2 — Per cluster**: brief synthesis when transitioning between topic clusters. "We've covered [area A] and [area B]. Before I ask about [area C], here is what I understood so far: [summary]. Does that capture it?" Confirms completeness, gives stakeholder a recovery point.
+- **Level 2 — Per group**: brief synthesis when transitioning between behavior groups. "We've covered [area A] and [area B]. Before I ask about [area C], here is what I understood so far: [summary]. Does that capture it?" Confirms completeness, gives stakeholder a recovery point.
 - **Level 3 — End of session**: full synthesis of everything discussed. Present to stakeholder for approval. This is the accuracy gate, the baseline signal, and the input to domain modeling.
 
 Do not introduce topic labels or categories during active listening. The summary must reflect what the stakeholder said, not new framing that prompts reactions to things they haven't considered.
@@ -98,20 +98,20 @@ Do not introduce topic labels or categories during active listening. The summary
 4. Run a **silent pre-mortem** on the confirmed synthesis: "Imagine we build exactly what was described, ship it, and it fails. What was missing?" Add any discoveries as new questions to the Questions table.
 5. Mark `Template §1: CONFIRMED` in `discovery.md`. This unlocks Session 2.
 
-### Session 2 — Cluster / Big Picture
+### Session 2 — Behavior Groups / Big Picture
 
-**Before the session**: Review the confirmed Session 1 synthesis. Identify topic clusters (cross-cutting concerns, system-wide constraints, integration points, lifecycle questions). Prepare cluster-level questions.
+**Before the session**: Review the confirmed Session 1 synthesis. Identify behavior groups (cross-cutting concerns, system-wide constraints, integration points, lifecycle questions). Prepare group-level questions.
 
-**During the session**: Apply Level 1 active listening per answer. Apply Level 2 active listening when transitioning between clusters. Apply CIT, Laddering, and CI Perspective Change per cluster. Add new questions in the moment.
+**During the session**: Apply Level 1 active listening per answer. Apply Level 2 active listening when transitioning between groups. Apply CIT, Laddering, and CI Perspective Change per group. Add new questions in the moment.
 
 **After the session**:
 
-1. For each cluster, write a **Cluster Summary** in `discovery.md`.
+1. For each group, write a **Group Summary** in `discovery.md`.
 2. Mark `Template §2: CONFIRMED` in `discovery.md`. This unlocks Session 3.
 
 ### Session 3 — Synthesis Approval + Feature Derivation
 
-**Before the session**: Produce a **Full Synthesis** across all clusters from Sessions 1 and 2. Write it to `discovery.md`.
+**Before the session**: Produce a **Full Synthesis** across all behavior groups from Sessions 1 and 2. Write it to `discovery.md`.
 
 **During the session**: Present the full synthesis to the stakeholder. "This is my understanding of the full scope. Please correct anything that is missing or wrong." Stakeholder approves or corrects. PO refines until the stakeholder explicitly approves.
 
@@ -152,26 +152,26 @@ Commit: `feat(discovery): baseline project discovery`
 3. Run a **silent pre-mortem** on the confirmed synthesis.
 4. Mark `Template §1: CONFIRMED`. This unlocks Session 2.
 
-### Session 2 — Cluster / Big Picture for This Feature
+### Session 2 — Behavior Groups / Big Picture for This Feature
 
-**Before the session**: Review the confirmed Session 1 synthesis. Identify clusters of behavior within this feature (happy paths, error paths, edge cases, lifecycle events, integration points).
+**Before the session**: Review the confirmed Session 1 synthesis. Identify behavior groups within this feature (happy paths, error paths, edge cases, lifecycle events, integration points).
 
-**During the session**: Apply Level 1 active listening per answer. Apply Level 2 active listening when transitioning between clusters. Apply CIT, Laddering, and CI Perspective Change per cluster.
+**During the session**: Apply Level 1 active listening per answer. Apply Level 2 active listening when transitioning between groups. Apply CIT, Laddering, and CI Perspective Change per group.
 
 **After the session**:
 
-1. Write **Cluster Summaries** in the `.feature` file. Name each cluster — these names become candidate `Rule:` titles.
+1. Write **Group Summaries** in the `.feature` file. Name each group — these names become candidate `Rule:` titles.
 2. Mark `Template §2: CONFIRMED`. This unlocks Session 3.
 
 ### Session 3 — Feature Synthesis Approval + Story Derivation
 
-**Before the session**: Produce a **Full Synthesis** of the feature scope, covering all clusters from Sessions 1 and 2.
+**Before the session**: Produce a **Full Synthesis** of the feature scope, covering all behavior groups from Sessions 1 and 2.
 
 **During the session**: Present the full synthesis to the stakeholder. Stakeholder approves or corrects. PO refines until explicitly approved.
 
 **After the session** (PO alone):
 
-1. Map each named cluster from Session 2 to a candidate user story (Rule).
+1. Map each named group from Session 2 to a candidate user story (Rule).
 2. Write `Status: BASELINED (YYYY-MM-DD)` to the `.feature` file's discovery section.
 3. Mark `Template §3: CONFIRMED`.
 
@@ -355,16 +355,16 @@ Feature: <Feature title>
   Template §1: CONFIRMED
   Synthesis: <PO synthesis — confirmed by stakeholder>
 
-  Session 2 — Cluster / Big Picture:
+  Session 2 — Behavior Groups / Big Picture:
   | ID | Question | Answer | Status |
   |----|----------|--------|--------|
   | Q2 | ... | ... | OPEN / ANSWERED |
   Template §2: CONFIRMED
-  Clusters:
-  - <Cluster name>: <one-sentence summary>
+  Behavior Groups:
+  - <Behavior group name>: <one-sentence summary>
 
   Session 3 — Feature Synthesis:
-  Synthesis: <full synthesis across all clusters>
+  Synthesis: <full synthesis across all behavior groups>
   Template §3: CONFIRMED — stakeholder approved YYYY-MM-DD
 
   Rule: <User story title>
@@ -409,19 +409,19 @@ Template §1: CONFIRMED
 Synthesis: <PO synthesis — confirmed by stakeholder>
 Pre-mortem: <gaps identified; new questions added above>
 
-## Session 2 — Cluster / Big Picture
+## Session 2 — Behavior Groups / Big Picture
 
 | ID | Question | Answer | Status |
 |----|----------|--------|--------|
 | Q2 | ... | ... | OPEN / ANSWERED |
 
 Template §2: CONFIRMED
-Clusters:
-- <Cluster name>: <one-sentence summary>
+Behavior Groups:
+- <Behavior group name>: <one-sentence summary>
 
 ## Session 3 — Full Synthesis
 
-<3–6 paragraph synthesis of all scope, clusters, and boundaries>
+<3–6 paragraph synthesis of all scope, behavior groups, and boundaries>
 
 Template §3: CONFIRMED — stakeholder approved YYYY-MM-DD
 ```

.opencode/skills/scope/discovery-template.md

@@ -20,13 +20,13 @@ Feature: <feature-name>
   Synthesis: (fill after stakeholder confirms)
   Pre-mortem: (fill after synthesis is confirmed)
 
-  Session 2 — Cluster / Big Picture:
+  Session 2 — Behavior Groups / Big Picture:
   | ID | Question | Answer | Status |
   |----|----------|--------|--------|
 
   Template §2: PENDING
-  Clusters:
-  - (fill after all cluster questions are answered)
+  Behavior Groups:
+  - (fill after all group questions are answered)
 
   Session 3 — Feature Synthesis:
   (fill after Sessions 1 and 2 are complete)

AGENTS.md

@@ -58,20 +58,20 @@ STEP 5: ACCEPT         (product-owner)  → demo, validate, move folder to compl
 PO creates `docs/features/discovery.md` using the 3-session template. **Skip Phase 1 entirely if `discovery.md` Status is BASELINED.** To add features to an existing project: append new questions to Session 1 and re-fill from there.
 
 - **Session 1** — Individual scope elicitation: 5Ws + Success + Failure + Out-of-scope. Gap-finding per answer using CIT, Laddering, and CI Perspective Change. PO writes synthesis; stakeholder confirms or corrects. PO runs silent pre-mortem on confirmed synthesis. Template §1 must be confirmed before Session 2.
-- **Session 2** — Cluster / big picture: questions target clusters and cross-cutting concerns. Gap-finding per cluster. Level 2 synthesis when transitioning between clusters. Template §2 must be complete before Session 3.
-- **Session 3** — Synthesis approval + feature derivation: PO produces full synthesis of all clusters; stakeholder approves or corrects (PO refines until approved). Domain analysis: nouns/verbs → subject areas → FDD "Action object" feature names. Create `backlog/<name>.feature` stubs. Write `Status: BASELINED` to `discovery.md`.
+- **Session 2** — Behavior groups / big picture: questions target behavior groups and cross-cutting concerns. Gap-finding per group. Level 2 synthesis when transitioning between groups. Template §2 must be complete before Session 3.
+- **Session 3** — Synthesis approval + feature derivation: PO produces full synthesis of all behavior groups; stakeholder approves or corrects (PO refines until approved). Domain analysis: nouns/verbs → subject areas → FDD "Action object" feature names. Create `backlog/<name>.feature` stubs. Write `Status: BASELINED` to `discovery.md`.
 
 ### Phase 2 — Feature Discovery (per feature)
 Each `.feature` file has its own 3-session discovery template in its description. **Sessions are enforced by the template: each section must be filled before proceeding to the next.**
 
 - **Session 1** — Individual entity elicitation: populate Entities table from project discovery; generate questions from entity gaps using CIT, Laddering, CI Perspective Change. PO writes synthesis; stakeholder confirms. Silent pre-mortem on confirmed synthesis.
-- **Session 2** — Cluster / big picture: questions target clusters of behavior within this feature. Gap-finding per cluster. Level 2 cluster transition summaries.
-- **Session 3** — Feature synthesis approval + story derivation: PO produces synthesis of feature scope and clusters; stakeholder approves or corrects (PO refines until approved). Clusters become candidate user stories (Rules). Write `Status: BASELINED` to `.feature` discovery section.
+- **Session 2** — Behavior groups / big picture: questions target behavior groups within this feature. Gap-finding per group. Level 2 group transition summaries.
+- **Session 3** — Feature synthesis approval + story derivation: PO produces synthesis of feature scope and behavior groups; stakeholder approves or corrects (PO refines until approved). Story candidates become candidate user stories (Rules). Write `Status: BASELINED` to `.feature` discovery section.
 
 **Decomposition check**: after Session 3, does this feature span >2 distinct concerns OR have >8 candidate Examples? YES → split into separate `.feature` files, re-run Phase 2. NO → proceed.
 
 ### Phase 3 — Stories (PO alone)
-Clusters from Phase 2 Session 2 → one `Rule:` block per user story. Each `Rule:` has the user story header (`As a / I want / So that`) as its description — no `Example:` blocks yet. INVEST gate: all 6 letters must pass. Commit: `feat(stories): write user stories for <name>`
+Story candidates from Phase 2 Session 2 → one `Rule:` block per user story. Each `Rule:` has the user story header (`As a / I want / So that`) as its description — no `Example:` blocks yet. INVEST gate: all 6 letters must pass. Commit: `feat(stories): write user stories for <name>`
 
 ### Phase 4 — Criteria (PO alone)
 Pre-mortem per Rule (all Rules must be checked before writing Examples). Write `Example:` blocks — declarative Given/When/Then, MoSCoW triage (Must/Should/Could) per Example. Review checklist (4.3). Commit: `feat(criteria): write acceptance criteria for <name>`

docs/features/discovery.md

@@ -23,14 +23,14 @@ Pre-mortem: (fill after synthesis is confirmed)
 
 ---
 
-## Session 2 — Cluster / Big Picture
+## Session 2 — Behavior Groups / Big Picture
 
 | ID | Question | Answer | Status |
 |----|----------|--------|--------|
 
 Template §2: PENDING
-Clusters:
-- (fill after all cluster questions are answered)
+Behavior Groups:
+- (fill after all group questions are answered)
 
 ---
 

docs/workflow.md

@@ -46,20 +46,20 @@ Each step has a designated agent and a specific deliverable. No step is skipped.
 │      → PO runs silent pre-mortem on confirmed synthesis             │
 │      [template §1: synthesis confirmed → unlocks Session 2]         │
 │                                                                     │
-│    Session 2 — Cluster / Big Picture                                │
-│      Questions target clusters and cross-cutting concerns           │
-│      Gap-finding per cluster: CIT · Laddering · CI Perspective      │
+│    Session 2 — Behavior Groups / Big Picture                       │
+│      Questions target behavior groups and cross-cutting concerns   │
+│      Gap-finding per group: CIT · Laddering · CI Perspective         │
 │      [new questions from elucidation added in the moment]           │
 │      Level 1: paraphrase each answer                                │
-│      Level 2: synthesis when transitioning between clusters         │
-│      [template §2: all clusters answered → unlocks Session 3]       │
+│      Level 2: synthesis when transitioning between groups        │
+│      [template §2: all groups answered → unlocks Session 3]          │
 │                                                                     │
 │    Session 3 — Synthesis Approval + Feature Derivation              │
-│      PO produces full synthesis across all clustered areas          │
+│      PO produces full synthesis across all behavior groups         │
 │      → stakeholder approves or corrects; PO refines until approved  │
 │      [template §3: approval → unlocks domain analysis]              │
 │      Domain analysis: nouns/verbs → subject areas                   │
-│      Name features (FDD "Action object" / Affinity clusters)        │
+│      Name features (FDD "Action object" / Affinity groups)           │
 │      Create backlog/<name>.feature stubs                            │
 │      Status: BASELINED written to discovery.md                      │
 │                                                                     │
@@ -75,18 +75,18 @@ Each step has a designated agent and a specific deliverable. No step is skipped.
 │      → PO runs silent pre-mortem on confirmed synthesis             │
 │      [template §1: synthesis confirmed → unlocks Session 2]         │
 │                                                                     │
-│    Session 2 — Cluster / Big Picture for this Feature               │
-│      Questions target clusters of behavior within this feature      │
-│      Gap-finding per cluster: CIT · Laddering · CI Perspective      │
+│    Session 2 — Behavior Groups / Big Picture for this Feature        │
+│      Questions target behavior groups within this feature            │
+│      Gap-finding per group: CIT · Laddering · CI Perspective         │
 │      [new questions from elucidation added in the moment]           │
-│      Level 1: paraphrase · Level 2: cluster transition summaries    │
-│      [template §2: all clusters answered → unlocks Session 3]       │
-│                                                                     │
-│    Session 3 — Feature Synthesis Approval + Story Derivation        │
-│      PO produces synthesis of feature scope and clusters            │
-│      → stakeholder approves or corrects; PO refines until approved  │
-│      Clusters → candidate user stories (Rules)                      │
-│      Status: BASELINED written to .feature discovery section        │
+│      Level 1: paraphrase · Level 2: group transition summaries      │
+│      [template §2: all groups answered → unlocks Session 3]          │
+│                                                                     │
+│    Session 3 — Feature Synthesis Approval + Story Derivation       │
+│      PO produces synthesis of feature scope and behavior groups     │
+│      → stakeholder approves or corrects; PO refines until approved │
+│      Story candidates → candidate user stories (Rules)               │
+│      Status: BASELINED written to .feature discovery section         │
 │      [template §3: approval + stories → unlocks decomp check]       │
 │                                                                     │
 │    DECOMPOSITION CHECK                                              │
@@ -95,7 +95,7 @@ Each step has a designated agent and a specific deliverable. No step is skipped.
 │      NO  → proceed                                                  │
 │                                                                     │
 │  Phase 3 — Stories (PO alone)                                       │
-│    Clusters from Phase 2 Session 2 → one Rule: block per story      │
+│    Story candidates from Phase 2 Session 2 → one Rule: block per story │
 │    INVEST gate: all 6 letters must pass before committing           │
 │    commit: feat(stories): write user stories for <name>             │
 │                                                                     │
@@ -370,12 +370,12 @@ Feature: <title>
   | ID | Question | Answer | Status |     ← OPEN / ANSWERED
   Synthesis: <PO synthesis — confirmed by stakeholder>
 
-  Session 2 — Cluster / Big Picture:
+  Session 2 — Behavior Groups / Big Picture:
   | ID | Question | Answer | Status |
-  Clusters: <named topic clusters derived from answers>
+  Behavior Groups: <named behavior groups derived from answers>
 
   Session 3 — Feature Synthesis:
-  Synthesis: <full synthesis across clusters>
+  Synthesis: <full synthesis across behavior groups>
   Approved: YES / NO
 
   Architecture:                         ← added at Step 2 by software-engineer