feat(github-automation): branch strategy integration with NATS observability

.claude/CLAUDE.md

@@ -213,16 +213,29 @@ PMOVES.AI is a **production-ready multi-agent orchestration platform** featuring
 - Services: DB (Postgres 17.6.1), GoTrue, PostgREST v14.3, Kong 3.7.1, Realtime v2.72.0, Storage v1.37.1, Studio, imgproxy, pg-meta, Edge Functions, Analytics (Logflare), Vector, Supavisor
 - Canonical consumer URL: `http://supabase-kong:8000/rest/v1` (via Kong gateway)
 - Standard variable names: `JWT_SECRET`, `ANON_KEY`, `SERVICE_ROLE_KEY` (SUPABASE_* aliases for compat)
+n**Neo4j** [Port 7474 HTTP, 7687 Bolt, 2004 Metrics]
+- Graph database for knowledge management, CHIT consciousness taxonomy, agent memory
+- Profile-based integration: make -C pmoves neo4j-local-up
+- API: POST http://localhost:7474/db/neo4j/tx/commit (Cypher transactions)
+- Health: GET http://localhost:7474/db/neo4j/health
+- **Use for:** Graph queries, relationship traversal, CHIT consciousness taxonomy, agent memory
+- **Submodule:** PMOVES-Neo4j (root-level, follows PMOVES-supabase pattern)
+- **See:** `PMOVES-Neo4j/CLAUDE.md` for submodule context
 - Schema: `pmoves_core`, Archon prompts
 - **Use for:** Metadata storage, content records, agent state, auth, connection pooling
 
 **Qdrant** [Port 6333]
 - Vector embeddings for semantic search
 - Collection: `pmoves_chunks`
 
-**Neo4j** [Port 7474 HTTP, 7687 Bolt]
-- Knowledge graph storage
-- Entity relationships, graph traversal
+**Neo4j** [Port 7474 HTTP, 7687 Bolt, 2004 Metrics]
+- Graph database for knowledge management, CHIT consciousness taxonomy, agent memory
+- Profile-based integration: make -C pmoves neo4j-local-up
+- API: POST http://localhost:7474/db/neo4j/tx/commit (Cypher transactions)
+- Health: GET http://localhost:7474/db/neo4j/health
+- **Use for:** Graph queries, relationship traversal, CHIT consciousness taxonomy, agent memory
+- **Submodule:** PMOVES-Neo4j (root-level, follows PMOVES-supabase pattern)
+- **See:** `PMOVES-Neo4j/CLAUDE.md` for submodule context
 
 **Meilisearch** [Port 7700]
 - Full-text keyword search

.claude/context/credentials-workflow.md

@@ -55,10 +55,18 @@ and available for runtime token minting by services (BoTZ MCP gateway, Archon).
 | Purpose | Ephemeral tokens for GHCR, cross-repo ops, MCP GitHub tools |
 | Action | `actions/create-github-app-token@v2` |
 | Workflows | `build-images.yml`, `integrations-ghcr.yml`, `self-hosted-builds.yml` |
-| Local setup | `docs/GITHUB_APP_LOCAL_SETUP.md` |
+| Quick start | `make github-app-setup` (automated) |
+| Docs | `pmoves/docs/GITHUB_APP_QUICK_START.md`, `pmoves/docs/AGENTS/GITHUB_APP_CREDENTIALS.md` |
 | Strategy doc | `pmoves/docs/infrastructure/github-app-strategy.md` |
 | Secrets | `GH_APP_ID`, `GH_APP_SEC` (PEM), `GH_APP_CLIENT_ID`, `GH_APP_INSTALLATION_ID` |
 
+**Automated Setup:**
+```bash
+cd pmoves
+make github-app-setup    # Uncomment credentials + run secrets-funnel
+make github-app-verify  # Verify all components
+```
+
 **Token hierarchy:**
 - **CI:** `actions/create-github-app-token@v2` → ephemeral GitHub App installation token for GHCR
 - **Runtime:** Services mint installation tokens from `GH_APP_ID` + `GH_APP_SEC`

.claude/context/evoswarm.md

@@ -73,11 +73,8 @@ NVML_ENABLED=true                                # Enable NVIDIA power monitorin
 **`GET /healthz`**
 - Health check for Evo Controller
 
-**`GET /swarm/status`**
-- Current swarm status, population info, best fitness
-
-**`POST /swarm/force-evolution`**
-- Manually trigger evolution cycle (for testing)
+**`GET /config`**
+- Current runtime configuration (`poll_seconds`, `sample_limit`, namespace, REST wiring)
 
 ## Genome Definition
 
@@ -332,7 +329,8 @@ CREATE TABLE geometry_swarm_runs (
 ### Check Controller Status
 
 ```bash
-curl http://localhost:8113/swarm/status
+curl http://localhost:8113/healthz
+curl http://localhost:8113/config
 ```
 
 Response:
@@ -379,11 +377,11 @@ uvicorn app:app --reload --port 8113
 ### Testing Evolution
 
 ```bash
-# Force evolution cycle
-curl -X POST http://localhost:8113/swarm/force-evolution
+# Check controller liveness
+curl http://localhost:8113/healthz
 
-# Check results
-curl http://localhost:8113/swarm/status
+# Inspect active controller configuration
+curl http://localhost:8113/config
 ```
 
 ### Integration Testing
@@ -428,7 +426,7 @@ Planned EvoSwarm features:
 **For Claude Code CLI users:**
 - EvoSwarm runs automatically in the background
 - Parameter packs are consumed transparently by geometry services
-- Monitor evolution via `/swarm/status` endpoint
+- Monitor controller health via `/healthz` and runtime settings via `/config`
 - Use test namespace for development: `namespace="test"`
 - Check CGP metadata to see which pack was used
 - Telemetry feeds back to controller automatically

.claude/context/geometry-nats-subjects.md

@@ -153,6 +153,13 @@ geometry.<type>.<event>.v1      # Core geometry events
   ```
 - **Subscribers:** publisher-discord, optimization dashboards
 
+### Credential Rotated
+
+**`tokenism.credential.rotated.v1`**
+- **Direction:** Published by ToKenism credential manager → Consumed by analytics, security
+- **Purpose:** Notification when a ToKenism credential (API key, signing key) is rotated
+- **Subscribers:** security dashboards, audit logs
+
 ### Geometry Event (Voice)
 
 **`tokenism.geometry.event.v1`**
@@ -206,6 +213,14 @@ geometry.<type>.<event>.v1      # Core geometry events
 - **Note:** Used for real-time geometry updates to connected clients
 - **Subscribers:** Hi-RAG v2 WebSocket consumers
 
+### CGP Calibration
+
+**`geometry.cgp.calibration.v1`**
+- **Direction:** Published by evo-controller → Consumed by Hi-RAG v2, analytics
+- **Purpose:** CGP calibration events from EvoSwarm RL parameter optimization
+- **Payload:** Calibration results with updated CGP parameters
+- **Subscribers:** Hi-RAG v2, parameter dashboards
+
 ### Geometry Event Raw
 
 **`geometry.event.v1`**