code-ai-installer 2.2.0 → 2.4.0

package/domains/product/locales/en/.agents/workflows/product-pipeline-rules.md

@@ -0,0 +1,340 @@
+---
+name: product-pipeline-rules
+description: Absolute rules of the product pipeline — gates, discipline, camp independence, inter-session protocol
+---
+
+# Product Pipeline Rules
+
+> [!CAUTION]
+> **🔴 ABSOLUTE RULE #1:** The pipeline cannot be skipped. No exceptions.
+> Every gate → read the agent → deliverable → `$gates` PASS → `$handoff` → **User sign-off** → next gate.
+>
+> **🔴 ABSOLUTE RULE #2:** In Full A/B — **Camp Independence**. Alpha and Beta do not see each other's work until Mediator. Violation = P0 contamination.
+
+---
+
+## Four Pipeline Modes
+
+### 🔵 Full Pipeline A — `/ship-right-thing` (6 sessions)
+
+Full strategic cycle with adversarial on the **Customer vs Business axis**. Mediator synthesizes. Output — PRD + Product Review Deck (PDF).
+
+```
+Session 1: CONDUCTOR → DISCOVERY                                        → session-1-handoff.md
+Session 2: CONDUCTOR → CUSTOMER-CHAMPION (product_strategist α)         → session-2-handoff.md
+Session 3: CONDUCTOR → BUSINESS-CHAMPION (product_strategist β)         → session-3-handoff.md
+Session 4: CONDUCTOR → MEDIATOR (Strategy Synthesis)                    → session-4-handoff.md
+Session 5: CONDUCTOR → PM → UX_DESIGNER + TECH_LEAD + DATA_ANALYST      → session-5-handoff.md
+Session 6: CONDUCTOR → DESIGNER → LAYOUTER → RELEASE_GATE               → PDF
+```
+
+### 🔵 Full Pipeline B — `/shape-prioritize` (6 sessions)
+
+Scope-decision cycle with adversarial on the **Build vs Cut axis**. Strategy is approved, scope is open. Mediator synthesizes the final MoSCoW. Output — PRD (PDF).
+
+```
+Session 1: CONDUCTOR → DISCOVERY (scope intake)                         → session-1-handoff.md
+Session 2: CONDUCTOR → BUILD-CAMP (pm α + tech_lead α)                  → session-2-handoff.md
+Session 3: CONDUCTOR → CUT-CAMP (pm β + tech_lead β)                    → session-3-handoff.md
+Session 4: CONDUCTOR → MEDIATOR (Scope Synthesis)                       → session-4-handoff.md
+Session 5: CONDUCTOR → PM (final PRD) → UX_DESIGNER + DATA_ANALYST      → session-5-handoff.md
+Session 6: CONDUCTOR → DESIGNER → LAYOUTER → RELEASE_GATE               → PDF
+```
+
+### 🟢 Spec — `/spec` (1 session)
+
+Strategy and scope are approved. A formal PRD is needed. No adversarial. Output — PRD (markdown).
+
+```
+CONDUCTOR → PM → UX_DESIGNER + TECH_LEAD → PRD (markdown)
+```
+
+### 🟡 Quick — `/quick-pm` (1 session)
+
+Quick assessment, focused question, low stakes. No adversarial. Output — Short brief (markdown).
+
+```
+CONDUCTOR → DISCOVERY → product_strategist → PM → TECH_LEAD → DATA_ANALYST → Short Brief
+```
+
+---
+
+## Decision Tree — Mode Selection
+
+```
+Strategy approved?
+  ├── NO + full plan needed                → /ship-right-thing (Full A)
+  ├── YES + scope unclear, scope debate    → /shape-prioritize (Full B)
+  ├── YES + scope approved, PRD needed     → /spec
+  └── Quick assessment, focused question  → /quick-pm
+
+Unsure? → 5 clarifying questions in Conductor Clarification Gate → propose mode with rationale
+```
+
+Escalation: if Conductor sees that the request does not match — asks 5 clarifying questions and proposes a mode with rationale. Transition only on "Approved".
+
+---
+
+## Mandatory Rules (7) — MANDATORY ENFORCEMENT
+
+### Rule 1: Gates Cannot Be Skipped
+- Gate order is defined by the workflow mode.
+- Each gate requires: deliverable + `$gates` PASS + `$handoff` + user sign-off.
+- Skipping a gate → 🔴 **P0 / BLOCKER**.
+
+**Enforcing skills:** `$gates` (readiness check), `$board` (status tracking), `$handoff` (context transfer).
+
+### Rule 2: Clarification First
+- In Session 1 — Conductor asks **at least 5 clarifying questions** before Discovery work begins.
+- Without answers to mandatory questions (mode / product task / segment / constraints / format) — P0 Blocker.
+
+**Enforcing skills:** `$gates` (COND-01 criteria include "5+ questions answered").
+
+### Rule 3: Handoff Is Mandatory Between Sessions
+- Without a handoff file, the next session does not start.
+- Handoff contains the **full text of artifacts** (not links — the next session has no context).
+- A ready-made prompt is generated (`$session-prompt-generator`) for Full A/B.
+
+**Enforcing skills:** `$handoff` (Session type), `$session-prompt-generator` (prompt), `$board` (state saving).
+
+### Rule 4: Adversarial Independence (Full A/B)
+- Camp Alpha and Camp Beta work **independently**, they do not see each other's artifacts until Mediator.
+- Beta receives the Alpha deliverable as a read-only input in Session 3 (via camp filter).
+- Alpha handoff to Beta — marked `team: beta` + filter rules.
+- Violation (contamination) → P0, Mediator marks `contamination: true`, confidence is lowered.
+
+**Enforcing skills:** `$handoff` (camp marker), `$gates` (camp independence check at COND-0N).
+
+### Rule 5: Evidence > Opinion
+- Every position in a deliverable — with a source (Discovery evidence inventory or external).
+- Without evidence → explicit mark `⚠️ assumed`.
+- Mediator accounts for this during scoring (lowering Evidence Strength).
+- PRD with ≥ 30% assumed positions without a risk register → P1 Gap.
+
+**Enforcing skills:** `$gates` (evidence coverage check), `$assumption-mapping` (risk × uncertainty).
+
+### Rule 6: Mediator Does Not Write Their Own Strategy
+- Mediator **only synthesizes** from the two available positions (Alpha + Beta).
+- Third way without grounding in the camps → P0 "Third Way Invention".
+- Synthesis paths: Adopt Alpha / Adopt Beta / Hybrid — always with source-pointer per element.
+
+**Enforcing skills:** `$gates` (MED-01 criteria include source-pointer verification).
+
+### Rule 7: Release Gate Is Final
+- PDF / PRD is not delivered to the user without `$gates` sign-off at RG-01.
+- All preceding gates must be [✓].
+- User sign-off received.
+
+**Enforcing skills:** `$gates` (RG-01 criteria), `$handoff` (final envelope).
+
+---
+
+## Mapping "Rule → Skill → Check"
+
+| Rule | Skills | What is checked | On violation |
+|------|--------|-----------------|--------------|
+| 1. Gates cannot be skipped | `$gates`, `$board`, `$handoff` | Board statuses are sequential, no gaps | 🔴 P0 Blocker |
+| 2. Clarification First | `$gates` (COND-01) | 5+ questions asked, answers received | 🔴 P0 Blocker |
+| 3. Handoff mandatory | `$handoff` (Session), `$session-prompt-generator` | File saved, artifacts as full text, prompt generated | 🔴 P0 Blocker |
+| 4. Camp Independence | `$handoff` (camp marker), `$gates` (contamination check) | Camp marker correct, no cross-camp reads | 🔴 P0 Blocker |
+| 5. Evidence > Opinion | `$gates` (evidence coverage), `$assumption-mapping` | Every position with source or `⚠️ assumed` mark | 🟠 P1 Gap (or P0 if coverage < 50%) |
+| 6. Mediator synthesis | `$gates` (MED-01 criteria), source-pointer verification | Unified deliverable with source per element | 🔴 P0 Blocker |
+| 7. Release Gate | `$gates` (RG-01), `$handoff` | All gates [✓], PDF tested, user sign-off | 🔴 P0 Blocker |
+
+---
+
+## Mechanical Blocks (enforced checks)
+
+| # | Block | When | What it checks | On failure |
+|---|-------|------|----------------|------------|
+| 1 | **Pre-flight check** | Before writing deliverable | Quote the user's last "Approved" | Stop, request Approved |
+| 2 | **Skill read check** | Before applying a skill | Read SKILL.md via `view_file` | Read, then apply |
+| 3 | **Camp marker check** | Full A/B Sessions 2-3 | Handoff contains `team: alpha` or `team: beta` | Reverse Handoff to Conductor |
+| 4 | **Evidence pointer check** | Every deliverable in Sessions 2-5 | Every position with source or `⚠️ assumed` | Mark or remove |
+| 5 | **Handoff file check** | Before closing session (Full A/B) | `docs/product/session-N-handoff.md` saved, self-contained | Do not close session, save file |
+| 6 | **Contamination check** | Mediator Session 4 | Alpha and Beta do not contain direct quotes of each other without marking | Mark contamination: true, lower confidence |
+| 7 | **PRD metrics check** | PM-SPEC (Session 5 / Spec) | PRD contains success metrics section (NSM / OKR link) | Do not pass to further gates |
+
+---
+
+## Gate Sequencing
+
+| Gate ID | Session | Agent | Deliverable | Severity on failure |
+|---------|:-------:|-------|-------------|---------------------|
+| COND-01 | 1 | Conductor | Interview Brief (5+ questions answered) | Blocker |
+| DISC-01 | 1 | Discovery | Discovery Brief (JTBD + problems + assumptions + evidence) | Blocker |
+| STRAT-01α | 2 (Full A) | product_strategist α | Strategy Brief Alpha (Customer-Champion) | Blocker |
+| STRAT-02β | 3 (Full A) | product_strategist β | Strategy Brief Beta (Business-Champion) | Blocker |
+| PM-01α | 2 (Full B) | PM α + Tech Lead α | Scope Proposal Alpha (Build-Camp) | Blocker |
+| PM-02β | 3 (Full B) | PM β + Tech Lead β | Scope Proposal Beta (Cut-Camp) | Blocker |
+| MED-01 | 4 (Full A/B) | Mediator | Unified Synthesis (Strategy or Scope) | Blocker |
+| PM-SPEC | 5 | PM | Final PRD + prioritized backlog | Blocker |
+| UX-01 | 5 | UX Designer | User flows + wireframes | Gap (PRD ok without for non-UI) |
+| TECH-01 | 5 | Tech Lead | Feasibility + NFR + risks + epic breakdown | Blocker |
+| DATA-01 | 5 | Data Analyst | Metric plan + hypotheses + experiments | Blocker |
+| DS-01 | 6 | Designer | Design spec (PRD layout + Deck) | Blocker |
+| LY-01 | 6 | Layouter | HTML self-contained, print-ready | Blocker |
+| RG-01 | 6 | Conductor | Release Gate sign-off | Blocker |
+
+## Severity Levels
+
+| Priority | Description | Action |
+|:--------:|-------------|--------|
+| 🔴 **P0 / BLOCKER** | Blocks progress / PDF publication | Immediately: fix or escalate |
+| 🟠 **P1 / IMPORTANT / GAP** | Important to fix before finalization, can continue with risk register | By Release Gate |
+| 🟡 **P2 / NICE-TO-HAVE / NOTE** | Can be included in the next iteration | Backlog |
+
+---
+
+## Escalation Protocol
+
+### When user sign-off is missing
+1. Show the deliverable and request "Approved".
+2. If no response — **wait**. Do not proceed to the next gate.
+3. Remind: "Waiting for 'Approved' to proceed to [next gate]".
+4. **Never** auto-proceed. `ShouldAutoProceed = false` ALWAYS.
+
+### On repeated FAIL at `$gates`
+| Iteration | Action |
+|:---------:|--------|
+| 1st FAIL | Standard return: list of gaps → agent revises |
+| 2nd FAIL | Extended return + specific instructions on "how to fix" |
+| 3+ FAIL | **Escalate to user** via `$gates` protocol (options: scope revision / return to previous gate / forced CONDITIONAL PASS) |
+
+### When a problem is discovered at a subsequent gate (Reverse Handoff)
+1. Agent creates a Reverse Handoff describing the problem via `$handoff`.
+2. Conductor routes the return to the appropriate gate.
+3. `$board` updates: current gate → `[↩]`, return gate → `[→]`.
+4. After fix — repeat `$gates`.
+
+### Stalemate (Mediator cannot synthesize)
+- Both camps are weak (low evidence, low coherence).
+- Mediator does not perform forced synthesis.
+- Escalate to user: two options + rationale + recommended path (usually — rerun one of the camps).
+
+### Contradictory evidence
+- Data from the two camps is incompatible.
+- Loop back to Discovery for additional interviews / data pull.
+- Or escalate to user for a solo-decision.
+
+### Compliance block
+- A requirement is discovered that needs legal / security review.
+- Pause pipeline, external review, return afterwards.
+
+### Resource block
+- Cannot be implemented within the given constraints.
+- Tech Lead escalates to Cut-Camp mode or requests a timeline / budget revision.
+
+---
+
+## Pipeline Health Metrics
+
+`$board` calculates and displays on every update:
+
+| Metric | Formula | Healthy | Problematic |
+|--------|---------|:-------:|:-----------:|
+| Progress (%) | Completed / Total × 100 | Growing | Stagnating |
+| Avg time/gate | Σ time / completed | ≤ 20 min | > 40 min |
+| Blockers (total) | Count of [!] for all time | 0-1 | ≥ 3 |
+| Returns (total) | Count of [↩] for all time | 0-1 | ≥ 3 |
+| FAIL rate | FAILs / gate checks × 100 | < 20% | > 40% |
+| Evidence coverage | % positions with ✅ Verified | ≥ 80% | < 50% |
+| Camp contamination | false / true | false | true |
+
+**When FAIL rate > 40%** or **contamination = true:** Conductor escalates to user — possibly, scope is too wide, data is insufficient, or the process has been violated.
+
+---
+
+## Pipeline Artifacts
+
+### Full Pipeline A (ship-right-thing)
+
+| Session | Artifact | Format | Created by | Skill |
+|:-------:|----------|--------|------------|-------|
+| 1 | Interview Brief | Markdown | Conductor | — |
+| 1 | Discovery Brief | Markdown (in handoff) | Discovery | `$jtbd-canvas`, `$problem-statement`, `$assumption-mapping` |
+| 1 | Session 1 Handoff | `docs/product/session-1-handoff.md` | Conductor | `$handoff`, `$session-prompt-generator` |
+| 2 | Strategy Brief Alpha (Customer-Champion) | Markdown | product_strategist α | `$product-vision`, `$north-star-metric`, `$okr-framework`, `$product-roadmap` |
+| 2 | Session 2 Handoff | `docs/product/session-2-handoff.md` | Conductor | `$handoff`, `$session-prompt-generator` |
+| 3 | Strategy Brief Beta (Business-Champion) | Markdown | product_strategist β | (same skills) |
+| 3 | Session 3 Handoff | `docs/product/session-3-handoff.md` | Conductor | `$handoff`, `$session-prompt-generator` |
+| 4 | Unified Strategy Synthesis | Markdown | Mediator | — |
+| 4 | Session 4 Handoff | `docs/product/session-4-handoff.md` | Conductor | `$handoff`, `$session-prompt-generator` |
+| 5 | Final PRD + backlog | Markdown | PM | `$prd-template`, `$user-story`, `$acceptance-criteria`, `$rice-scoring` |
+| 5 | UX output (flows + wireframes) | Markdown | UX Designer | `$user-flow`, `$design-brief` |
+| 5 | Tech Brief (full feasibility) | Markdown | Tech Lead | `$epic-breakdown` |
+| 5 | Metric Plan | Markdown | Data Analyst | `$saas-metrics`, `$hypothesis-template`, `$ab-test-design`, `$aarrr-metrics` |
+| 5 | Session 5 Handoff | `docs/product/session-5-handoff.md` | Conductor | `$handoff`, `$session-prompt-generator` |
+| 6 | Design Spec | Markdown | Designer | `$report-design` |
+| 6 | PRD HTML + Product Review Deck HTML | HTML → PDF | Layouter | `$html-pdf-report` |
+| 6 | Release Gate sign-off | — | Conductor | `$gates` |
+
+### Full Pipeline B (shape-prioritize)
+
+| Session | Artifact | Format | Created by | Skill |
+|:-------:|----------|--------|------------|-------|
+| 1 | Discovery Brief (scope intake) | Markdown | Discovery | `$jtbd-canvas`, `$assumption-mapping` |
+| 2 | Scope Proposal Alpha (Build-Camp) | Markdown | PM α + Tech Lead α | `$rice-scoring`, `$epic-breakdown` |
+| 3 | Scope Proposal Beta (Cut-Camp) | Markdown | PM β + Tech Lead β | `$moscow-prioritization`, `$epic-breakdown` |
+| 4 | Unified Scope Decision | Markdown | Mediator | — |
+| 5 | Final PRD + UX + Metric plan (same as Full A S5, without new strategy) | Markdown | PM, UX, Data Analyst | (see Full A) |
+| 6 | PRD HTML + mini deck → PDF | HTML → PDF | Designer + Layouter | `$report-design`, `$html-pdf-report` |
+
+### Spec
+
+| Artifact | Format | Created by | Skill |
+|----------|--------|------------|-------|
+| PRD | Markdown | PM | `$prd-template`, `$user-story`, `$acceptance-criteria` |
+| UX notes (if UI) | Markdown | UX Designer | `$user-flow` |
+| Tech notes | Markdown | Tech Lead | `$epic-breakdown` |
+
+### Quick
+
+| Artifact | Format | Created by | Skill |
+|----------|--------|------------|-------|
+| Short Brief | Markdown (1-2 pages) | PM (with contributions from others) | `$rice-scoring` (rough), `$saas-metrics` |
+
+---
+
+## Localization
+
+- Default: `ru`
+- Each artifact can be localized via `locales/en/` (Phase 2, not included in Phase 1).
+
+---
+
+## Pipeline Launch Checklist (for Conductor)
+
+Conductor performs on initialization:
+
+- [ ] Mode determined by Decision Tree (Full A / Full B / Spec / Quick)
+- [ ] 5+ clarifying questions asked and answered
+- [ ] `$board` created with correct gate IDs (from Gate Sequencing)
+- [ ] First gate activated (COND-01 → [→])
+- [ ] User sign-off received on mode and scope
+- [ ] For a continuing session: handoff file loaded and verified (recovery protocol `$handoff`)
+
+## Session Close Checklist (for Conductor)
+
+Conductor performs before wrapping up:
+
+- [ ] All gates of the current session are [✓] or [⊘] (no [→] or [!])
+- [ ] `$board` saved in the handoff file
+- [ ] All artifacts included as **full text** in handoff
+- [ ] Camp marker set correctly (Full A/B Sessions 2-3)
+- [ ] Evidence coverage checked (≥ 80% or risk register updated)
+- [ ] `$session-prompt-generator` called, prompt generated (Full A/B)
+- [ ] Handoff file saved: `docs/product/session-N-handoff.md`
+- [ ] File is self-contained (test: work can start without external context)
+- [ ] User sign-off received for session close
+
+## Release Gate Checklist (for Conductor, Session 6)
+
+- [ ] All gates of all sessions are [✓] on `$board`
+- [ ] PRD HTML / PDF ready (LY-01 self-containment check passed)
+- [ ] Product Review Deck ready (Full A)
+- [ ] Visual check: page breaks clean, cross-links work, TOC correct
+- [ ] File size reasonable (< 10 MB PRD / < 5 MB deck)
+- [ ] Content intact — not modified by Layouter
+- [ ] User sign-off received
+- [ ] Final board + log shown to user

package/domains/product/locales/en/.agents/workflows/quick-pm.md

@@ -0,0 +1,292 @@
+# /quick-pm — Quick Pipeline (1 session, no adversarial)
+
+> **Use for quick product answers — 1-2 hours of work.**
+> 1 session, no adversarial, no handoff files between steps. Output — Short brief (markdown, 1-2 pages).
+
+## When to Use
+
+- Focused product question with obvious scope
+- No time for Full Pipeline (6 sessions)
+- Stakes are low — decision is reversible or low-cost
+- An assessment is needed, not a full plan
+- Unlock decision: "should we X?" / "is Y worth building?"
+- Triage ticket / backlog item before adding to quarter plan
+
+## When NOT to Use
+
+- Strategic decision (pivot, new segment, $M investment) → `/ship-right-thing`
+- Scope debate (team disagrees) → `/shape-prioritize`
+- A formal PRD is needed for engineering → `/spec`
+- High stakes (quarters of work, irreversible) → `/ship-right-thing`
+- Compliance-sensitive decision → Full A/B for thorough risk review
+
+## Decision Tree (clarifying questions in Step 0)
+
+```
+Stakes?
+  ├── Low (reversible, < $10K impact) → /quick-pm (this workflow)
+  ├── Medium ($10-100K, team-quarter impact) → /spec or /shape-prioritize
+  └── High (quarters, pivot, > $100K) → /ship-right-thing
+
+Time budget?
+  ├── 1-2 hours → /quick-pm
+  ├── 3-6 hours → /spec
+  └── Multi-day → Full A/B
+
+Scope clarity?
+  ├── Obvious (one-pager) → /quick-pm
+  ├── Mostly clear, needs spec → /spec
+  └── Open debate → /shape-prioritize
+
+Output type?
+  ├── Short brief / recommendation → /quick-pm
+  ├── Engineering-ready PRD → /spec
+  └── Exec-review deck → /ship-right-thing
+```
+
+## Pipeline (1 session)
+
+```
+CONDUCTOR → DISCOVERY → product_strategist → PM → TECH_LEAD → DATA_ANALYST → Short Brief
+```
+
+All agents in one session, sequentially, without handoff files between steps. Each agent — compact contribution.
+
+## Input
+
+| Field | Required | Description |
+|-------|:--------:|-------------|
+| Focused question | ✅ | Format: "Should we X?" or "How do we prioritize Y?" in 1 sentence |
+| Current context | ✅ | ARR, segment, recent decisions — 2-3 sentences |
+| Time box | ⬚ | Default 1-2 hours; max 3 hours |
+| Existing evidence (if any) | ⬚ | Support tickets, user feedback, metrics |
+| Decision deadline | ⬚ | If present — limits depth |
+
+## Steps
+
+### Step 0 — Clarification (5 min)
+
+**Conductor**: 3-5 clarifying questions (not 5+ as in Full):
+- What exactly are we deciding (is it reversible?)
+- Stakes (cost of mistake)?
+- Time box?
+- Answer format (recommendation / estimate / comparison)?
+- Any hard constraints?
+
+### Step 1 — Discovery Compact (15-20 min)
+
+**Discovery**:
+- JTBD summary (1 paragraph for end-user + 1 for buyer if applicable)
+- Problem statement (1 sentence)
+- Key assumptions (3-5)
+- Evidence pointer (if available — what data already confirms the direction)
+
+Format: 30-60 lines markdown.
+
+### Step 2 — Strategy Alignment (15 min)
+
+**product_strategist**:
+- Which NSM / OKR is linked (reference to approved strategy)
+- Which roadmap theme
+- Strategic fit? (strongly / loosely / off-strategy)
+- If off-strategy → escalate to /ship-right-thing
+
+Format: 20-40 lines markdown.
+
+### Step 3 — PM Assessment (20 min)
+
+**PM**:
+- Rough RICE scoring per item (top 3-5 candidates)
+- MVP sketch — 3-5 user stories (no Gherkin AC)
+- Kano classification (must / performance / delighter)
+- Out-of-scope (brief)
+
+Format: 40-80 lines markdown.
+
+### Step 4 — Tech Sanity (10 min)
+
+**Tech Lead**:
+- Feasibility rough (S/M/L/XL) overall
+- Top 2-3 risks
+- Hidden integrations / dependencies? (Yes / No, if Yes — highlight)
+- Spike recommendation if unknown unknowns
+
+Format: 20-30 lines markdown.
+
+### Step 5 — Metric Sketch (10 min)
+
+**Data Analyst**:
+- Primary success metric (aligned with approved NSM)
+- Guardrail (what must not break — churn / NPS / perf)
+- Rough target (number + timeframe)
+
+Format: 10-20 lines markdown.
+
+### Step 6 — Short Brief (final deliverable, 5 min)
+
+**PM** consolidates:
+
+```markdown
+# Quick Brief — [Question]
+
+## Recommendation
+[Yes / No / Need more discovery] — 1 sentence
+
+## Rationale (top 3 reasons)
+1. ...
+2. ...
+3. ...
+
+## Rough Scope (if proceed)
+- [bullet 1]
+- [bullet 2]
+- [bullet 3]
+
+## Success Metric
+- Primary: [NSM link]
+- Guardrail: [what must not break]
+- Rough target: [number + timeframe]
+
+## Key Risks
+- [risk 1 + mitigation]
+- [risk 2 + mitigation]
+
+## Effort Estimate
+- T-shirt: S / M / L
+- Team-weeks: ~N
+
+## Next Step
+- [Usually: deeper dive via Full A/B, proceed to /spec, or decline]
+```
+
+Format: 1-2 pages markdown.
+
+## Gate sequencing (simplified)
+
+Within one session — all gates in order:
+
+```
+COND-01 → DISC-quick → STRAT-quick → PM-quick → TECH-quick → DATA-quick → LITE-RG
+```
+
+Each gate: deliverable + `$gates` mini-check + next step. No handoff files. User sign-off — on the **final brief**, not on each step.
+
+## Severity levels
+
+- **P0:** missing context (ARR, segment, product area)
+- **P1:** weak evidence (acceptable — explicitly "preliminary", with noted assumption map)
+- **P2:** incomplete strategic fit analysis
+
+## Escalation paths
+
+If Quick reveals complexity — **convert to Full A or Full B**:
+
+- **Discovery reveal:** problem space is more complex than 3-5 assumptions → save brief, escalate to `/ship-right-thing`
+- **Scope debate emerges:** team disagrees on scope within the session → escalate to `/shape-prioritize`
+- **Strategy misalignment:** initiative off-strategy / contradicts approved → escalate to `/ship-right-thing`
+- **High risk revealed:** Tech Lead surfaces compliance / security concern → pause, do not ship Quick brief as decision
+- **Overrun time box:** session > 3h → stop, save progress, escalate (Quick is not suitable for this scope)
+
+## Health metrics
+
+| Metric | Healthy | Problematic | Action |
+|--------|:-------:|:-----------:|--------|
+| Session duration | 1-2h | > 3h | Escalate to Full |
+| Open questions | ≤ 3 | > 6 | Scope too big for Quick |
+| Evidence coverage | ≥ 60% | < 30% | Discovery loop recommend, Quick brief as preliminary |
+| Strategic fit | Strong / Loose | Off-strategy | Escalate to /ship-right-thing |
+
+## Output
+
+- `quick-brief-[slug].md` (1-2 pages) — final deliverable
+- Optional: CONTEXT.md update (if project uses CONTEXT.md)
+- Optional: adding item to roadmap backlog (if recommendation = Yes)
+
+## Example — TeamFlow AI 1:1: Add emoji reactions to summary? (Quick)
+
+**Request:** "Should we add emoji reactions (👍/❤️/🎉) to the summary card? Support tickets are asking for it."
+
+**Mode:** /quick-pm (low stakes, focused, 1h)
+
+**Step 0 (Clarification):**
+- Stakes: low (1 week implementation, reversible).
+- Deadline: next sprint planning (2 days).
+- Format: recommendation + rough estimate.
+- Constraints: existing design system, accessibility.
+
+**Step 1 (Discovery, 15 min):**
+- End-user JTBD: emotional — "I want to react quickly to the summary without going back to the meeting".
+- Evidence: 6 support tickets (Q3-Q4) asking for quick reactions. Manager Slack feedback.
+- Assumptions: emoji reactions increase engagement; not distracting.
+
+**Step 2 (Strategy, 10 min):**
+- NSM link: sessions/week per active manager (engagement proxy). Roadmap theme: "delighters Q2".
+- Fit: loosely (not core JTBD, delighter category).
+
+**Step 3 (PM, 15 min):**
+- RICE: reach 1500 (MAM), impact 0.25 (minor), confidence 80%, effort 1 week → score 300.
+- 3 user stories (view reactions, add reaction, see who reacted).
+- Kano: delighter.
+
+**Step 4 (Tech, 10 min):**
+- Feasibility: Yes, S. Existing reactions service (extend).
+- Risks: accessibility (screen reader announces), perf (aggregation at 100+ reactions).
+- Integrations: existing.
+
+**Step 5 (Data, 5 min):**
+- Primary: reaction rate per summary (new event).
+- Guardrail: summary rating stays ≥ 4.0.
+- Target: 30% summaries get ≥ 1 reaction within 4 weeks.
+
+**Step 6 (Brief):**
+
+```markdown
+# Quick Brief — Emoji Reactions in Summary
+
+## Recommendation
+**Yes, add as delighter in Q2.**
+
+## Rationale
+1. 6 support tickets + Manager Slack direct ask (demand signal)
+2. 1 team-week effort, existing stack leverage
+3. Delighter category — differentiator without core risk
+
+## Rough Scope
+- 3 emojis (👍/❤️/🎉) in summary card
+- Reaction aggregation (count + who)
+- Notification to 1:1 participants (optional)
+
+## Success Metric
+- Primary: 30% summaries with ≥ 1 reaction in 4 weeks
+- Guardrail: summary rating stays ≥ 4.0
+- Rough target: 1500 MAM × 30% = 450 reactions/week
+
+## Key Risks
+- Accessibility: screen reader announces (mitigation: aria-label per reaction)
+- Perf at 100+ reactions (mitigation: cached aggregation, unlikely at current scale)
+
+## Effort Estimate
+- T-shirt: S (1 team-week)
+- 1 senior engineer
+
+## Next Step
+- Add to Q2 roadmap as delighter
+- No Full pipeline needed — proceed to /spec in sprint planning
+```
+
+**Result:** 1 session × 60 min = recommendation + rough plan, ready for sprint planning.
+
+---
+
+## Anti-patterns
+
+| Mistake | Why it's bad | How to do it right |
+|---------|-------------|--------------------|
+| Formal PRD in Quick | Quick = brief, not PRD | If a PRD is needed, escalate to /spec |
+| Adversarial in Quick | Adversarial is only for Full A/B | Single perspective ok for low stakes |
+| Overthinking / overrun | Defeats purpose of Quick | Boxed to 1-2h; > 3h → escalate |
+| Skip strategic fit | Off-strategy items slip in | Always check NSM / roadmap link |
+| No guardrail metric | NSM can be gamed | Guardrail — churn / NPS / perf always |
+| Decision without evidence | Guessing in recommendation | Include existing support tickets / metrics / user quotes |
+| Quick brief as PRD | Engineering expects more detail | Explicit "this is preliminary, PRD to follow" |
+| Missing next step | Brief without action | Always define next step (proceed / escalate / decline) |