Add Nora — AI Work Intelligence power

nora/POWER.md

@@ -0,0 +1,164 @@
+---
+name: "nora"
+displayName: "Nora — AI Work Intelligence"
+description: "Your AI gets smarter every session. Nora captures sessions, learns patterns, injects context, and compounds your AI Leverage — all locally."
+keywords: ["nora", "kernora", "patterns", "context", "memory", "intelligence", "learning", "steering", "session", "decisions", "bugs", "anti-patterns", "knowledge", "leverage", "coaching"]
+author: "Kernora AI"
+---
+
+# Nora — AI Work Intelligence
+
+Nora is your silent coding partner. She captures every session, extracts patterns, decisions, and anti-patterns, then feeds them back as context and steering. Your AI Leverage Score compounds toward 5.0x as Nora learns your patterns.
+
+100% local. Zero bytes leave your machine. No API key required on Mac.
+
+## Onboarding
+
+### Step 1: Check Prerequisites
+
+Verify the following are installed:
+
+- **Python 3.9+** — Run `python3 --version` to confirm
+- **git** — Run `git --version` to confirm
+
+If either is missing, halt setup and inform the user.
+
+### Step 2: Install Nora
+
+Nora bootstraps automatically when installed as a Kiro Power.
+
+For manual install:
+```bash
+git clone https://github.com/kernora-ai/nora.git /tmp/nora-install && bash /tmp/nora-install/install.sh && rm -rf /tmp/nora-install
+```
+
+### Step 3: Configure LLM (Optional)
+
+Nora works out of the box with no API key on Mac (Apple FoundationModels on macOS 26+, MLX-LM on macOS 14+).
+
+For cloud analysis, set any of these environment variables:
+- `ANTHROPIC_API_KEY`
+- `GEMINI_API_KEY`
+- `OPENAI_API_KEY`
+- `AWS_PROFILE` (for Bedrock)
+
+Provider priority: IDE LLM → Apple FoundationModels → MLX-LM → BYOK API keys → Ollama.
+
+### Step 4: Verify Installation
+
+1. Check dashboard: `curl -sf http://localhost:2742/health` (should return JSON with `"status":"ok"`)
+2. Check steering: `ls ~/.kiro/steering/nora-*.md` (should list 3 files)
+
+## Available MCP Server
+
+### nora
+
+**Connection:** Local stdio process
+
+Nora's MCP server provides 18 tools for querying your session intelligence. All data stays local in `~/.kernora/echo.db`.
+
+**Tools (18):**
+
+- **nora_search** — Full-text search across patterns, decisions, bugs, and insights
+- **nora_patterns** — List effective coding patterns, optionally filtered by project
+- **nora_decisions** — List architectural decisions recorded across sessions
+- **nora_bugs** — List known bugs with severity, file path, and fix code
+- **nora_stats** — Dashboard stats: sessions, insights, AI Leverage Score
+- **nora_session** — Get full details for a specific session by ID
+- **nora_scope_validation** — Validate execution scope before multi-file edits
+- **nora_skills** — Fetch distilled methodology from your best sessions
+- **nora_scan** — Seed database from git history
+- **nora_pe_review** — Principal Engineer 4-tier code audit
+- **nora_coe** — Blameless root cause investigation (5 whys)
+- **nora_coe_product** — Product COE — why was this built wrong
+- **nora_retro** — Engineering retrospective with git velocity
+- **nora_sofac** — Factory health status
+- **nora_inventory** — Feature audit: SHIP/POLISH/WIRE/BLOCKER
+- **nora_coach** — AI leverage coaching
+- **nora_onboard** — Onboard a new developer
+- **nora_help** — List all tools
+
+## How Nora Works
+
+1. **You code normally.** Nora is silent during your session.
+2. **Session ends.** Hooks capture the transcript locally.
+3. **Nora analyzes.** Extracts patterns, decisions, bugs, and your AI Leverage Score.
+4. **Steering evolves.** Steering files regenerate with the latest intelligence.
+5. **Next session is smarter.** Context from past sessions is injected into your prompts.
+
+## AI Leverage Score
+
+A composite metric measuring your AI effectiveness:
+
+| Score | Label | What it means |
+|-------|-------|--------------|
+| 1.0–2.0 | Early | AI isn't helping much yet |
+| 2.0–3.0 | Developing | Getting value, room to grow |
+| 3.0–4.0 | Strong | Measurably effective AI usage |
+| 4.0–5.0 | Excellent | Elite AI collaboration |
+
+Export your score as a shareable certificate from the Coach tab in the dashboard.
+
+## Privacy
+
+- Session transcripts stored in `~/.kernora/echo.db` on your machine
+- Local LLM analysis via Apple FoundationModels or MLX-LM — zero network calls
+- Optional cloud analysis uses YOUR API key if configured
+- Zero telemetry, zero cloud storage, zero data sharing
+
+## Dashboard
+
+Open http://localhost:2742 to see:
+
+| Tab | What it shows |
+|-----|--------------|
+| **Home** | AI Leverage Score, loop health, top projects, rule suggestions |
+| **Projects** | Per-project AI metrics, patterns, decisions, bugs |
+| **Activity** | Session history with outcome indicators |
+| **Coach** | AI Leverage sparkline, coaching notes, decision patterns, certificate |
+| **Knowledge** | Best practices, playbooks, anti-patterns |
+| **Memory** | Context injection feed, steering file viewer |
+| **Decisions** | Searchable architectural decisions |
+| **Bugs** | Bug inventory with severity, fix suggestions, mark resolved |
+| **Settings** | LLM provider config, local AI status |
+
+## Configuration
+
+Edit `~/.kernora/config.toml`:
+
+```toml
+[mode]
+type = "byok"           # your data stays on your machine
+
+[model]
+provider = "auto"       # tries IDE → local → BYOK → Ollama
+
+[dashboard]
+port = 2742
+```
+
+## Troubleshooting
+
+**Dashboard not loading?** Run `python3 ~/.kernora/app/dashboard.py` and check for errors. Most common: missing Flask dependency — run `pip3 install flask`.
+
+**Steering files empty?** Steering regenerates after the first analyzed session. Run a coding session, then check `~/.kiro/steering/nora-*.md`.
+
+**Analysis not running?** Check the Settings tab in the dashboard. Verify your LLM provider is connected (green indicator). On Mac with no API key, ensure macOS 26+ for Apple FoundationModels or macOS 14+ for MLX-LM.
+
+## Resources
+
+- [GitHub: kernora-ai/nora](https://github.com/kernora-ai/nora)
+- [Dashboard](http://localhost:2742)
+- [kernora.ai](https://kernora.ai)
+
+## License and support
+
+This power is licensed under Elastic License 2.0 (Elastic-2.0).
+
+**Privacy Policy**
+
+https://kernora.ai/privacy
+
+**Support**
+
+https://github.com/kernora-ai/nora/issues | hello@kernora.ai

nora/mcp.json

@@ -0,0 +1,9 @@
+{
+  "mcpServers": {
+    "nora": {
+      "command": "python3",
+      "args": ["${HOME}/.kernora/app/nora_mcp.py"],
+      "env": {}
+    }
+  }
+}

nora/steering/nora-antipatterns.md

@@ -0,0 +1,15 @@
+# Nora — Anti-Patterns
+
+> Auto-updated after session analysis. Activates automatically in Kiro.
+
+Anti-patterns are mistakes, failed approaches, and bugs that Nora has observed.
+They act as guardrails — Nora warns you before you repeat a known mistake.
+
+## How anti-patterns work
+
+1. Nora identifies bugs, failed approaches, and error patterns from sessions
+2. Each anti-pattern includes the error signature, affected files, and suggested fix
+3. Anti-patterns are injected as warnings when Nora detects you're about to repeat one
+4. Mark anti-patterns as resolved in the Bugs tab when the underlying issue is fixed
+
+*This file populates after your first analyzed session.*

nora/steering/nora-decisions.md

@@ -0,0 +1,15 @@
+# Nora — Architectural Decisions
+
+> Auto-updated after session analysis. Activates automatically in Kiro.
+
+Decisions record what was chosen, why, and what alternatives were considered.
+Nora extracts these from session transcripts so your team never re-debates settled questions.
+
+## How decisions work
+
+1. Nora identifies architectural and design choices from session transcripts
+2. Each decision includes rationale, alternatives considered, and affected files
+3. Decisions are injected into future sessions when a similar choice arises
+4. Prevents the same debate from recurring across sessions or team members
+
+*This file populates after your first analyzed session.*

nora/steering/nora-patterns.md

@@ -0,0 +1,15 @@
+# Nora — Effective Patterns
+
+> Auto-updated after session analysis. Activates automatically in Kiro.
+
+Patterns are extracted from your coding sessions and ranked by effectiveness score (0–1.0).
+Higher scores mean the pattern is frequently used and consistently produces good outcomes.
+
+## How patterns work
+
+1. Nora observes tool usage and code changes during each session
+2. After analysis, recurring effective approaches are extracted as patterns
+3. Patterns are injected into future sessions when relevant context matches
+4. Effectiveness scores update as Nora observes whether injected patterns are adopted
+
+*This file populates after your first analyzed session.*