Remove remaining weftbox/weft-llm references from configs\n\nGenericized infrastructure-specific names in agent config comments:\n- 'weftbox' → 'the server' (6 config files + threat model)\n- 'weft-llm' → 'llm-service' (prometheus scrape target name)\n- 'weft-llm-gpu' → 'llm-gpu' (grafana dashboard UID example)\n\nZero 'weft' references remain in any tracked file (except 'goweft'\nGitHub org name). Configs re-signed. 126 tests passing.

Author: goweft

agents/grafana-bridge.yaml

@@ -1,5 +1,5 @@
 # Heddle Agent: grafana-bridge
-# Bridges the Grafana API on weftbox (port 3005).
+# Bridges the Grafana API on the server (port 3005).
 # Auth via credential broker (grafana-basic-auth).
 # 1 dashboard (WEFT LLM & GPU Metrics), 1 datasource (Prometheus), 4 alert rules.
 
@@ -27,7 +27,7 @@ agent:
       parameters:
         uid:
           type: string
-          description: "Dashboard UID, e.g. 'weft-llm-gpu'"
+          description: "Dashboard UID, e.g. 'llm-service-gpu'"
           required: true
       returns:
         type: string

agents/nexus-bridge.yaml

@@ -1,5 +1,5 @@
 # Heddle Agent: nexus-bridge
-# Bridges the NEXUS AI Platform API on weftbox (port 8080).
+# Bridges the NEXUS AI Platform API on the server (port 8080).
 # 45 endpoints available, exposing the most useful ones for Claude.
 # NEXUS provides multi-LLM routing, RAG, code analysis, and app management.
 

agents/ollama-bridge.yaml

@@ -1,5 +1,5 @@
 # Heddle Agent: ollama-bridge
-# Bridges the Ollama API on weftbox (port 11434).
+# Bridges the Ollama API on the server (port 11434).
 # 7 models installed, GPU-accelerated inference.
 
 agent:

agents/prometheus-bridge.yaml

@@ -1,14 +1,14 @@
 # Heddle Agent: prometheus-bridge
-# Bridges the Prometheus API on weftbox (port 9092).
-# 4 scrape targets: node-exporter, prometheus, intel-rag, weft-llm
+# Bridges the Prometheus API on the server (port 9092).
+# 4 scrape targets: node-exporter, prometheus, intel-rag, llm-service
 
 agent:
   name: prometheus-bridge
   version: "1.0.0"
   description: >
     Bridge to the Prometheus monitoring API. Query metrics, check
     target health, list alerts, and run PromQL queries across
-    weftbox infrastructure (node-exporter, intel-rag, weft-llm).
+    server infrastructure (node-exporter, intel-rag, llm-service).
 
   model:
     provider: none

agents/vram-orchestrator.yaml

@@ -10,7 +10,7 @@ agent:
   name: vram-orchestrator
   version: "1.0.0"
   description: >
-    GPU VRAM orchestrator for weftbox. Manages model loading across
+    GPU VRAM orchestrator for server. Manages model loading across
     Ollama and llama.cpp, tracks VRAM usage, provides intelligent
     model selection, and optimizes GPU memory allocation.
     24GB RX 7900 XTX with 30 GGUFs + 7 Ollama models available.

docs/threat-model.md

@@ -206,7 +206,7 @@ Credential Policy:
 **Now implemented (Phase 3f):**
 - **Generated agent quarantine.** AI-generated configs land in a staging directory and require explicit promotion before they can be registered or run.
 - **Config signing.** YAML configs can be cryptographically signed; the runtime verifies signatures before loading.
-- **URL allowlist.** Bridge URLs are checked against a set of known weftbox service endpoints.
+- **URL allowlist.** Bridge URLs are checked against a set of known server service endpoints.
 
 **Framework mapping:**
 | Control | OWASP Agentic | NIST AI RMF | MAESTRO |