@lucapimenta/copy-chief-black 1.0.0

package/framework/commands/dag.md

@@ -0,0 +1,196 @@
+# /dag — DAG Pipeline Visualization & Validation
+
+> Sprint S40.8 — DAG Pipeline Engine
+> Visualize and inspect any pipeline as a Directed Acyclic Graph.
+> State persists across sessions alongside each DAG file.
+
+## Usage
+
+```
+/dag <subcommand> [options]
+```
+
+## Subcommands
+
+### /dag validate <pipeline>
+
+Validate a DAG pipeline for structural integrity.
+
+Checks: no cycles, all edge references valid, gate nodes have thresholds, no orphan nodes.
+
+```
+/dag validate helix
+/dag validate production
+/dag validate research
+```
+
+**Implementation:**
+```bash
+bun run ~/copywriting-ecosystem/scripts/dag-executor.ts ~/.claude/workflows/<pipeline>-dag.yaml --validate
+```
+
+---
+
+### /dag status <pipeline>
+
+Show current status of all nodes in a pipeline.
+
+Displays nodes in topological order with status icons:
+- ✅ completed
+- 🔄 in_progress
+- ⬜ pending
+- 🔴 blocked (shows which deps are blocking)
+- ⏭️ skipped
+
+Also shows: summary counts, critical path.
+
+```
+/dag status helix
+/dag status production
+/dag status research
+```
+
+**Implementation:**
+```bash
+bun run ~/copywriting-ecosystem/scripts/dag-executor.ts ~/.claude/workflows/<pipeline>-dag.yaml --status
+```
+
+---
+
+### /dag next <pipeline>
+
+Show next executable nodes — nodes whose all dependencies are satisfied and that are ready to start.
+
+```
+/dag next helix
+/dag next production
+/dag next research
+```
+
+**Implementation:**
+```bash
+bun run ~/copywriting-ecosystem/scripts/dag-executor.ts ~/.claude/workflows/<pipeline>-dag.yaml --next
+```
+
+---
+
+### /dag critical-path <pipeline>
+
+Show the critical path through the pipeline — the longest chain of sequential dependencies.
+
+Useful for understanding the minimum time to complete the pipeline and which nodes can't be parallelized.
+
+```
+/dag critical-path helix
+/dag critical-path production
+/dag critical-path research
+```
+
+**Implementation:**
+```bash
+bun run ~/copywriting-ecosystem/scripts/dag-executor.ts ~/.claude/workflows/<pipeline>-dag.yaml --critical-path
+```
+
+---
+
+### /dag complete <pipeline> <node-id>
+
+Mark a node as completed and persist the state.
+
+Use this after completing a phase/task to unlock downstream nodes.
+
+```
+/dag complete helix research-gate
+/dag complete helix fase-01
+/dag complete helix fase-02
+/dag complete helix mup-gate
+```
+
+**Implementation:**
+```bash
+bun run ~/copywriting-ecosystem/scripts/dag-executor.ts ~/.claude/workflows/<pipeline>-dag.yaml --complete <node-id>
+```
+
+---
+
+### /dag reset <pipeline> <node-id>
+
+Reset a node back to pending status. Useful when re-doing a phase.
+
+```
+/dag reset helix fase-05
+/dag reset helix mup-gate
+```
+
+**Implementation:**
+```bash
+bun run ~/copywriting-ecosystem/scripts/dag-executor.ts ~/.claude/workflows/<pipeline>-dag.yaml --reset <node-id>
+```
+
+---
+
+## Pipeline Names
+
+| Name | DAG File | Description |
+|------|----------|-------------|
+| `helix` | `~/.claude/workflows/helix-dag.yaml` | HELIX 10-phase briefing pipeline |
+| `production` | `~/.claude/workflows/production-dag.yaml` | Copy production (VSL, LP, Criativos, Emails) |
+| `research` | `~/.claude/workflows/research-dag.yaml` | VOC Squad + competitor + mechanism research |
+
+## Common Node IDs
+
+### HELIX Pipeline
+- `research-gate` — entry gate (must be PASSED first)
+- `fase-01` through `fase-10` — HELIX phases
+- `mup-gate` — MUP validation (after fase-05)
+- `mecanismo-gate` — Mecanismo Único validation (after fase-06)
+- `briefing-gate` — exit gate
+
+### Research Pipeline
+- `context-load` — load niche library + CONTEXT.md
+- `voc-squad` — parallel group (5 analysts)
+- `voc-youtube`, `voc-instagram`, `voc-tiktok`, `voc-reddit`, `voc-amazon` — individual analysts
+- `competitor-spy` — ads library spy
+- `mechanism-research` — scientific backing
+- `avatar-synthesis` — avatar profile
+- `research-synthesis` — consolidated synthesis
+- `research-gate` — exit gate
+
+### Production Pipeline
+- `production-gate` — entry gate
+- `vsl-track`, `lp-track`, `creatives-track`, `emails-track` — parallel tracks
+- `vsl-chapter-1` through `vsl-chapter-6` — VSL chapters
+- `vsl-black-validation`, `lp-black-validation`, `creatives-black-validation`, `emails-black-validation` — final gates
+- `delivery-gate` — exit gate
+
+## State Persistence
+
+Node states are saved alongside each DAG file:
+- `~/.claude/workflows/helix-dag.state.json`
+- `~/.claude/workflows/production-dag.state.json`
+- `~/.claude/workflows/research-dag.state.json`
+
+State persists across sessions. Use `--reset` to undo a `--complete`.
+
+## Integration with Session Start
+
+At the beginning of each session, run `/dag status helix` to see where you are in the pipeline. This replaces manual inspection of `helix-state.yaml` for quick orientation.
+
+## Typical Workflow
+
+```bash
+# 1. Check pipeline state at session start
+/dag status helix
+
+# 2. See what to do next
+/dag next helix
+
+# 3. Complete a phase after finishing it
+/dag complete helix fase-01
+
+# 4. Validate DAG structure (when debugging)
+/dag validate helix
+
+# 5. See critical path for planning
+/dag critical-path helix
+```

package/framework/commands/debate.md

@@ -0,0 +1,232 @@
+---
+description: Structured debate between Copy Squad experts on a topic
+argument-hint: "[expert1] vs [expert2] [vs expert3] [vs expert4] sobre [topico] [--framework adversarial|socratic|collaborative]"
+---
+
+# Copy Squad Debate (N-Expert + Scoring)
+
+**Arguments:** $ARGUMENTS
+
+## Step 1: Parse Arguments
+
+Extract from `$ARGUMENTS`:
+- **experts**: names separated by "vs" (case-insensitive). Minimum 2, maximum 4.
+- **topic**: everything after "sobre" (or "about", "on", ":")
+- **framework**: if `--framework` flag present, extract value. Default: `adversarial`
+
+**Defaults:**
+- If only ONE expert name given (no "vs"), set expert2 = `copy-chief`
+- If NO expert names given, set expert1 = `halbert`, expert2 = `schwartz`
+- If no topic detected, ASK the user before proceeding
+- If no `--framework`, default to `adversarial`
+
+**Valid expert names** (match case-insensitive, use lowercase for file):
+abraham, agora, bencivenga, brown, brunson, caples, carlton, chaperon, cialdini, collier, copy-chief, halbert, hopkins, hormozi, kennedy, kern, makepeace, ogilvy, powers, reeves, ry-schwartz, sabri, schwartz, sugarman, vaynerchuk
+
+If an expert name is not in this list, report the error and list valid names. Do NOT proceed.
+
+**Framework options:**
+- `adversarial` (default) — experts challenge and attack each other's positions
+- `socratic` — experts ask provocative questions to each other, drawing out deeper thinking
+- `collaborative` — experts build on each other's ideas, seeking synthesis
+
+## Step 2: Load Voice DNA Profiles
+
+Read ALL expert profiles:
+- `~/.claude/copy-squad/{expert}.md` for each expert
+
+Extract from each file:
+- **Core Philosophy** section (worldview and principles)
+- **Voice Patterns** section (how they write and argue)
+- **Core Identity** line from the header
+
+These define HOW each expert speaks during the debate. Each expert MUST stay in character throughout all rounds.
+
+## Step 3: Detect Active Offer (Optional)
+
+Check if the current working directory contains an active offer:
+```bash
+find ~/copywriting-ecosystem -maxdepth 4 -name "CONTEXT.md" -newer ~/copywriting-ecosystem/.git/HEAD 2>/dev/null | head -3
+```
+
+If an active offer is detected, note its path for saving output later. If not, output will go to chat only.
+
+## Step 4: Run the Debate
+
+### Round Structure (varies by number of experts)
+
+**For 2 experts:** 5 rounds (Open, Counter, Rebuttal x2, Synthesis)
+**For 3 experts:** 7 rounds (Open x3, Counter-Cross x3, Synthesis)
+**For 4 experts:** 9 rounds (Open x4, Counter-Cross x4, Synthesis)
+
+**Format rules for ALL rounds:**
+- Each expert speaks IN CHARACTER using their Voice DNA patterns
+- 100-200 words per round (strict)
+- Use the expert's actual rhetorical devices, sentence structures, and terminology
+- Reference the expert's known frameworks and principles where relevant
+- The topic must be addressed CONCRETELY, not abstractly
+
+### Framework-Specific Behavior
+
+**Adversarial (default):**
+- Experts directly challenge opposing positions
+- Each rebuttal must identify a WEAKNESS in the other's argument
+- Steel-man before attacking (never strawman)
+
+**Socratic:**
+- Experts ask penetrating questions instead of making declarations
+- Each round ends with a question that forces the next expert to think deeper
+- Goal is to REVEAL insights through questioning, not win
+
+**Collaborative:**
+- Each expert builds on the previous expert's contribution
+- "Yes, AND..." structure — acknowledge, then extend
+- Goal is to create a synthesis GREATER than any individual perspective
+
+---
+
+### Opening Rounds — Each Expert Opens
+
+For each expert (in order), write:
+
+**Header:** `## Round {N}: {Expert} — Opening Position`
+
+{Expert} presents their approach to the topic. Must include:
+- Their core principle applied to this specific topic
+- A concrete recommendation or technique
+- Why their approach works (from their philosophical framework)
+
+Write in {Expert}'s voice using their documented patterns.
+
+---
+
+### Counter-Cross Rounds — Each Expert Responds
+
+For each expert (in order), write:
+
+**Header:** `## Round {N}: {Expert} — {Framework-verb}`
+
+Framework-verb:
+- Adversarial: "Counter-Attack"
+- Socratic: "Cross-Examination"
+- Collaborative: "Building On"
+
+Each expert responds to the PREVIOUS expert(s). Must include:
+- Framework-appropriate engagement with other positions
+- Refinement of their own position
+- Concrete examples or techniques
+
+---
+
+### Scoring Round (NEW)
+
+**Header:** `## Scoring Round: Cross-Expert Evaluation`
+
+After all debate rounds, each expert scores EVERY OTHER expert's arguments on a 1-10 scale:
+
+For each expert:
+```
+### {Expert}'s Scores:
+- {Other Expert 1}: X/10 — "{brief justification}"
+- {Other Expert 2}: X/10 — "{brief justification}"
+[... for each other expert]
+```
+
+Then compile:
+
+```
+### Score Matrix
+| Scorer → | Expert1 | Expert2 | Expert3 | Expert4 |
+|----------|---------|---------|---------|---------|
+| Expert1  | —       | X       | X       | X       |
+| Expert2  | X       | —       | X       | X       |
+| Expert3  | X       | X       | —       | X       |
+| Expert4  | X       | X       | X       | —       |
+| **AVG**  | X.X     | X.X     | X.X     | X.X     |
+
+**Winner: {Expert with highest average}** (X.X/10)
+```
+
+---
+
+### Synthesis Round
+
+**Header:** `## Synthesis — The Best of All Minds`
+
+This round is written by Copy Chief BLACK (not any expert). Must include:
+
+1. **Core Agreement**: What ALL experts actually agree on
+2. **Key Tensions**: The real philosophical differences that drive disagreements
+3. **Points of Consensus**: What emerged as irreconcilable vs what converged
+4. **Merged Recommendation**: A concrete, ACTIONABLE approach that takes the strongest elements from each side
+5. **When to Use Each**: Specific scenarios where each expert's approach wins
+6. **Applied to Topic**: 3-5 bullet points of what to DO based on this synthesis
+
+The synthesis MUST be specific enough to act on immediately. No abstractions. No "it depends." Concrete steps.
+
+---
+
+## Step 5: Output
+
+### Format the complete debate as:
+
+```markdown
+# Debate: {Expert1} vs {Expert2} [vs Expert3] [vs Expert4]
+## Topic: {topic}
+## Framework: {framework}
+Date: {YYYY-MM-DD}
+
+---
+
+{Opening Rounds}
+
+---
+
+{Counter-Cross Rounds}
+
+---
+
+{Scoring Round}
+
+---
+
+{Synthesis Round}
+
+---
+
+## Next Step
+
+Validate the synthesis using `consensus` (zen MCP) for multi-model agreement.
+This confirms the merged recommendation holds across different reasoning approaches.
+```
+
+### Save location:
+
+**If active offer detected:**
+```bash
+mkdir -p {offer-path}/production/debates/
+# Write to: {offer-path}/production/debates/debate-{experts-joined-by-vs}-{YYYY-MM-DD}.md
+```
+
+Report the file path to the user.
+
+**If no active offer:**
+Output the full debate in chat. Do NOT create files in arbitrary locations.
+
+## Constraints
+
+- NEVER break character during debate rounds. Each expert speaks ONLY in their documented voice.
+- NEVER let the synthesis be vague or abstract. It must produce actionable steps.
+- NEVER strawman any position. Steel-man ALL sides.
+- If the topic relates to an active offer, use offer-specific context (MUP, MUS, avatar) to ground the debate in reality.
+- Each round: 100-200 words. Not 50. Not 300. This is strict.
+- Scoring must include justification for EVERY score given.
+- With 3+ experts, ensure EACH expert gets equal speaking time.
+
+## Post-Debate Recommendation
+
+After the debate output, suggest:
+
+> To validate this synthesis with multi-model agreement, run `consensus` (zen MCP)
+> with the merged recommendation as input. This stress-tests the conclusion across
+> different reasoning models before applying it to production copy.

package/framework/commands/diagnose-offer.md

@@ -0,0 +1,184 @@
+---
+description: Diagnose why an offer isn't converting — cross-check all quality areas
+argument-hint: "[offer-name]"
+---
+
+# Diagnose Offer — Cross-System Quality Audit
+
+**Arguments:** $ARGUMENTS
+
+## Step 1: Locate the Offer
+
+Extract offer name from `$ARGUMENTS`. Search for the offer directory:
+
+```bash
+find ~/copywriting-ecosystem -maxdepth 3 -name "helix-state.yaml" 2>/dev/null
+```
+
+Match the offer name (case-insensitive, partial match OK) to find the offer root path.
+
+If no match found, list available offers and ask the user to clarify.
+
+## Step 2: Run 5 Diagnostic Checks
+
+For each check, assign a status: **PASS** (green), **WARN** (yellow), or **FAIL** (red).
+
+---
+
+### Check 1: VOC Quality
+
+Read `{offer}/research/synthesis.md` and `{offer}/research/voc/summary.md`.
+
+| Criterion | PASS | WARN | FAIL |
+|-----------|------|------|------|
+| synthesis.md exists | Yes | — | No |
+| Confidence score | >= 80% | 70-79% | < 70% or missing |
+| Platforms with quotes | 4+ | 2-3 | 0-1 |
+| Total quotes | 100+ | 50-99 | < 50 |
+| DRE defined | Yes, with escalada | Mentioned but vague | Missing |
+| Emotions categorized | 5+ categories | 3-4 | < 3 |
+
+**Score:** Count PASS=2pts, WARN=1pt, FAIL=0pts. Max 12. Normalize to /10.
+
+---
+
+### Check 2: MUP/MUS Quality
+
+Read `{offer}/mecanismo-unico.yaml`.
+
+| Criterion | PASS | WARN | FAIL |
+|-----------|------|------|------|
+| File exists | Yes | — | No |
+| State | APPROVED | VALIDATED | DRAFT/UNDEFINED |
+| MUP nova_causa filled | Yes, specific | Generic | Empty |
+| MUP sexy_cause filled | Yes, transmissible | Exists but weak | Empty |
+| MUS gimmick_name filled | Yes, sticky | Exists but weak | Empty |
+| MUS origin_story filled | Yes, with paradox | Exists but generic | Empty |
+| MUS authority_hook filled | Yes, super structure | Exists but weak | Empty |
+| RMBC avg score | >= 8 | 7-7.9 | < 7 or missing |
+
+**Score:** Count PASS=2pts, WARN=1pt, FAIL=0pts. Max 16. Normalize to /10.
+
+---
+
+### Check 3: Copy Quality (Production Gates)
+
+Read `{offer}/helix-state.yaml` for tools_used in production phase. Check for validation scores.
+
+| Criterion | PASS | WARN | FAIL |
+|-----------|------|------|------|
+| blind_critic used | Yes, score >= 8 | Used, score < 8 | Not used |
+| emotional_stress_test used | Yes, score >= 8 | Used, score < 8 | Not used |
+| layered_review used | Yes | — | Not used |
+| black_validation used | Yes, score >= 8 | Used, score < 8 | Not used |
+
+Also check for production files:
+
+| Criterion | PASS | WARN | FAIL |
+|-----------|------|------|------|
+| Production files exist | Expected count met | Partial | 0 files |
+
+**Score:** Normalize to /10.
+
+---
+
+### Check 4: Avatar Alignment
+
+Read `{offer}/research/avatar/summary.md` and `{offer}/briefings/phases/fase-02-*.md` or `fase-03-*.md`.
+
+| Criterion | PASS | WARN | FAIL |
+|-----------|------|------|------|
+| Avatar summary exists | Yes | — | No |
+| DRE explicitly defined | Yes, with type + escalada | Type only | Missing |
+| Escalada documented (5 levels) | 3+ levels | 1-2 levels | None |
+| Consciousness level defined | Yes | — | No |
+| Failed solutions listed | 4+ with quotes | 1-3 | None |
+| VOC quotes support DRE | 10+ quotes match | 5-9 | < 5 |
+
+**Score:** Normalize to /10.
+
+---
+
+### Check 5: Funnel Completeness
+
+Check production directories for expected deliverables.
+
+**For TSL (Landing Page) offers:**
+
+| Criterion | PASS | WARN | FAIL |
+|-----------|------|------|------|
+| LP blocks produced | 14/14 | 7-13 | < 7 |
+| Criativos produced | 3+ | 1-2 | 0 |
+| Email sequence | Exists | — | Missing |
+
+**For VSL offers:**
+
+| Criterion | PASS | WARN | FAIL |
+|-----------|------|------|------|
+| VSL chapters produced | 6/6 | 3-5 | < 3 |
+| LP blocks produced | 14/14 | 7-13 | < 7 |
+| Criativos produced | 3+ | 1-2 | 0 |
+| Email sequence | Exists | — | Missing |
+
+**Score:** Normalize to /10.
+
+---
+
+## Step 3: Generate Diagnostic Report
+
+### Output Format
+
+```markdown
+# Diagnostic Report: {offer-name}
+## Date: {YYYY-MM-DD}
+
+---
+
+### Summary
+
+| Area | Score | Status |
+|------|-------|--------|
+| VOC Quality | X/10 | PASS/WARN/FAIL |
+| MUP/MUS Quality | X/10 | PASS/WARN/FAIL |
+| Copy Quality | X/10 | PASS/WARN/FAIL |
+| Avatar Alignment | X/10 | PASS/WARN/FAIL |
+| Funnel Completeness | X/10 | PASS/WARN/FAIL |
+| **Overall** | **X/10** | **STATUS** |
+
+Status thresholds: >= 8 = PASS, 6-7.9 = WARN, < 6 = FAIL
+
+---
+
+### Probable Root Cause
+
+{Identify the area with the LOWEST score. Explain WHY this is likely the bottleneck.}
+
+### Recommended Actions (by priority)
+
+1. **[CRITICAL]** {Action for lowest-scoring area}
+2. **[HIGH]** {Action for second-lowest area}
+3. **[MEDIUM]** {Actions for WARN areas}
+
+### Detailed Findings
+
+{For each of the 5 checks, list every criterion with its status and specific evidence found.}
+```
+
+## Step 4: Save Report
+
+**If the offer has a production/ directory:**
+```bash
+mkdir -p {offer-path}/reviews/
+# Write to: {offer-path}/reviews/diagnostic-{YYYY-MM-DD}.md
+```
+
+**If no production/ directory:**
+Output the full report in chat.
+
+## Constraints
+
+- NEVER guess scores. Read the actual files and count real evidence.
+- NEVER skip a check. All 5 areas must be evaluated even if files are missing (that's a FAIL).
+- The root cause analysis must be SPECIFIC, not "needs improvement." Say exactly WHAT is missing.
+- Recommended actions must be ACTIONABLE — specific commands or files to create/edit.
+- If an offer is in early phases (research/briefing), it's EXPECTED that production checks fail. Note this context.