refactor: modular 80/20 skill architecture isolation

skills/pareto-audit/README.md

@@ -0,0 +1,124 @@
+# pareto-audit — AI-Powered 80/20 Workflow Audit
+
+**Skill ID:** `pareto-audit` | **Version:** 1.0.0 | **Complexity:** 6.5 | **Reproducibility:** 0.81
+
+Identifies the 20% of inputs driving 80% of outputs across any domain. Produces ranked action matrix, pruning recommendations, and unconventional reframes via the Pareto Cognitive Engine (PCE).
+
+---
+
+## Quick Start
+
+```bash
+# Deploy to Cloudflare
+chmod +x deploy.sh
+./deploy.sh
+
+# Set API key
+npx wrangler secret put ANTHROPIC_API_KEY
+```
+
+**Live URL:** `https://pareto.genesisconductor.io`
+
+---
+
+## API
+
+```bash
+POST https://pareto.genesisconductor.io/api/audit
+Content-Type: application/json
+
+{
+  "domain": "startup",
+  "goal": "Reach $10K MRR in 90 days",
+  "activities": ["Cold outreach", "Discovery calls", "Feature building"],
+  "outputs": "3 customers, $450 MRR"
+}
+```
+
+### Input Schema
+
+| Field | Type | Required | Description |
+|---|---|---|---|
+| `domain` | string | Yes | Context: startup, engineering, product, marketing, career, workflow |
+| `goal` | string | Yes | Primary objective with measurable target |
+| `activities` | array | Yes | Current activities (min 3, max 20) |
+| `outputs` | string | No | Observed metrics/results |
+
+### Output Schema
+
+```json
+{
+  "summary": "Core insight in one sentence",
+  "vital_20": [{ "activity", "contribution_pct", "insight" }],
+  "trivial_80": [{ "activity", "contribution_pct", "prune_reason" }],
+  "unconventional_reframes": [{ "conventional_assumption", "pareto_reframe" }],
+  "action_matrix": [{ "rank", "action", "expected_impact", "effort" }],
+  "recursive_note": "When to re-run audit"
+}
+```
+
+---
+
+## Examples
+
+Validated outputs in `examples/` — ground truth corpus:
+
+| File | Domain | Key Finding |
+|---|---|---|
+| [`startup-founder-2026-03.json`](examples/startup-founder-2026-03.json) | Startup | Inbound conversion + discovery = 75% of output |
+| [`engineering-team-2026-03.json`](examples/engineering-team-2026-03.json) | Engineering | PR cycle time is the single throughput constraint |
+| [`product-manager-2026-03.json`](examples/product-manager-2026-03.json) | Product | 40% churn from incomplete onboarding |
+
+---
+
+## Architecture
+
+```
+pareto-audit/
+├── SKILL.md              # multi-skill-orchestrator-mcp registry definition
+├── README.md             # This file
+├── src/
+│   └── worker.js         # Cloudflare Worker (frontend + API)
+├── wrangler.jsonc        # Wrangler config
+├── deploy.sh             # One-command deployment
+└── examples/
+    ├── startup-founder-2026-03.json
+    ├── engineering-team-2026-03.json
+    └── product-manager-2026-03.json
+```
+
+---
+
+## Integration
+
+| Surface | Method |
+|---|---|
+| `multi-skill-orchestrator-mcp` | Skill ID: `pareto_audit` |
+| Genesis Conductor Coalition | Plugin gateway |
+| QMCP | MCP tool surface |
+| Direct HTTP | `POST /api/audit` |
+| Slack | `@pareto-audit` mention |
+
+---
+
+## Rate Limits
+
+| Tier | Price | Limit |
+|---|---|---|
+| Free | $0 | 3 audits/day/IP |
+| Pro | $19/mo | Unlimited |
+| Team | $149/mo | Team comparisons + export |
+
+---
+
+## PCE Reasoning Modes
+
+1. **Reflective** — Examines assumptions in input
+2. **Unconventional** — Inverts standard optimization logic
+3. **Non-linear** — Identifies compounding effects
+4. **Strategic** — Evaluates timing and sequencing
+5. **Hedonistic** — Considers sustainable effort allocation
+
+---
+
+*Part of the Genesis Conductor thermodynamic AI orchestration ecosystem.*

skills/pareto-audit/SKILL.md

@@ -0,0 +1,177 @@
+---
+name: pareto-audit
+version: 1.0.0
+description: >
+  Applies 80/20 Pareto analysis to any domain. Identifies the vital 20% of inputs
+  driving 80% of outputs. Produces ranked action matrix, pruning recommendations,
+  and unconventional reframes via the Pareto Cognitive Engine (PCE). Reasoning modes:
+  Reflective, Unconventional, Non-linear, Strategic, Hedonistic.
+complexity: 6.5
+reproducibility: 0.81
+tokens_estimate: 3200
+dependencies: []
+revenue_model: "free:3/day/IP | pro:$19/mo | team:$149/mo"
+deployment: "https://pareto.genesisconductor.io"
+---
+
+## Trigger Patterns
+
+Activation phrases for `multi-skill-orchestrator-mcp`:
+
+- "run a pareto audit on..."
+- "apply 80/20 analysis to..."
+- "identify the vital 20%..."
+- "what should I prune from..."
+- "pareto my workflow"
+- "find the leverage points in..."
+- "what's the 80/20 of..."
+
+---
+
+## Input Schema
+
+```yaml
+domain:
+  type: string
+  required: true
+  description: Business domain context (startup, engineering, product, marketing, career, workflow)
+
+goal:
+  type: string
+  required: true
+  description: Primary objective with measurable target and timeframe
+
+activities:
+  type: array[string]
+  required: true
+  min_items: 3
+  max_items: 20
+  description: Current activities consuming time/resources
+
+outputs:
+  type: string
+  required: false
+  description: Observed metrics, results, or baseline data (improves analysis quality)
+```
+
+---
+
+## Execution Protocol
+
+```
+State 1: INPUT VALIDATION
+  └─ Validate required fields (domain, goal, activities)
+  └─ Normalize activity list (deduplicate, standardize format)
+  └─ Extract implicit metrics from outputs field
+        ↓
+State 2: CONTRIBUTION MAPPING
+  └─ For each activity, estimate contribution to goal (0-100%)
+  └─ Apply Pareto heuristics: identify natural 80/20 clustering
+  └─ Flag activities with <5% contribution as trivial
+        ↓
+State 3: VITAL 20% EXTRACTION
+  └─ Select top contributors summing to ~80% output
+  └─ Generate insight explaining WHY each is vital
+  └─ Identify hidden dependencies between vital activities
+        ↓
+State 4: TRIVIAL 80% ANALYSIS
+  └─ List remaining activities
+  └─ Generate prune_reason for each (specific, actionable)
+  └─ Distinguish between prune (stop entirely) and defer (pause temporarily)
+        ↓
+State 5: UNCONVENTIONAL REFRAMES
+  └─ Identify 3 conventional assumptions in the input
+  └─ Generate Pareto-inverted perspectives
+  └─ Apply reasoning modes: Reflective, Unconventional, Non-linear
+        ↓
+State 6: ACTION MATRIX GENERATION
+  └─ Rank actions by (expected_impact × 1/effort)
+  └─ Include exactly 5 actions
+  └─ First action must be achievable within 24 hours
+        ↓
+State 7: RECURSIVE NOTE
+  └─ Identify when to re-run audit (milestone or time-based)
+  └─ Preview what the next Pareto layer will likely reveal
+```
+
+---
+
+## Output Schema (JSON)
+
+```json
+{
+  "summary": "string — One sentence capturing the core insight",
+
+  "vital_20": [
+    {
+      "activity": "string",
+      "contribution_pct": "number (0-100)",
+      "insight": "string — Why this activity matters disproportionately"
+    }
+  ],
+
+  "trivial_80": [
+    {
+      "activity": "string",
+      "contribution_pct": "number (0-100)",
+      "prune_reason": "string — Specific, actionable reason to stop or defer"
+    }
+  ],
+
+  "unconventional_reframes": [
+    {
+      "conventional_assumption": "string",
+      "pareto_reframe": "string — Inverted perspective applying 80/20 logic"
+    }
+  ],
+
+  "action_matrix": [
+    {
+      "rank": "number (1-5)",
+      "action": "string — Specific, actionable step",
+      "expected_impact": "enum: high | medium | low",
+      "effort": "enum: high | medium | low"
+    }
+  ],
+
+  "recursive_note": "string — When and why to re-run this audit"
+}
+```
+
+---
+
+## Integration Points
+
+| Surface | Method | Identifier |
+|---|---|---|
+| `multi-skill-orchestrator-mcp` | npm package | `pareto_audit` |
+| Genesis Conductor Coalition | Plugin gateway | `coalition-gateway` entitlement |
+| QMCP | MCP tool surface | `mcp.genesisconductor.io` |
+| Direct HTTP | REST API | `POST /api/audit` |
+| Slack Integration | Bot mention | `@pareto-audit` |
+
+---
+
+## PCE Reasoning Modes
+
+The Pareto Cognitive Engine applies five reasoning modes during analysis:
+
+1. **Reflective** — Examines assumptions embedded in the input
+2. **Unconventional** — Inverts standard optimization logic
+3. **Non-linear** — Identifies compounding effects and second-order consequences
+4. **Strategic** — Evaluates timing and sequencing of actions
+5. **Hedonistic** — Considers sustainable effort allocation and energy management
+
+---
+
+## Rate Limits
+
+| Tier | Audits | Features |
+|---|---|---|
+| Free | 3/day/IP | Standard analysis |
+| Pro ($19/mo) | Unlimited | + Export, history, saved templates |
+| Team ($149/mo) | Unlimited | + Team comparisons, aggregated insights |
+
+---
+
+*Part of the Genesis Conductor thermodynamic AI orchestration ecosystem.*