fix: align contracts to 4-name LiteLLM registry (2026-09-12)

- Remove retired names (qwen3.6-27B-code, qwen3.6-35B-udq4) from live alias claims
- Update Strix Halo model to Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf (strix-moe, 256K ctx)
- Fix litellm-health step 7 probe to gpu-vision (monitor key scoped)
- Move qwen3.6-27B-code/35B-udq4 from raw-but-live to non-resolving in audit
- Fold in pm2-self-heal: remove spoton-service (live PM2 set is 4/4)
- Update hermes templates, key enforcement, timeout tables to live names

audit-hermes-config.py

@@ -245,11 +245,10 @@ def audit(path):
         "gemma-4-12b": "gpu-vision",
         "crew-auto": "syslog-auto (its 64K cap is retired; no cap in force)",
         "ornith-1.0-35b": "strix-moe",
-    }
-    raw_but_live = {
         "qwen3.6-27B-code": "gpu-dense",
         "qwen3.6-35B-udq4": "strix-moe",
     }
+    raw_but_live = {}
     for field_path, value in _iter_model_values(cfg):
         if value in non_resolving:
             check(

gpu-fleet.prose.md

@@ -91,7 +91,7 @@ Single source of truth for models, aliases, rpm caps, weights and fallback chain
 CT 116 `/opt/inference-harness/litellm_config.yaml`. Do not duplicate those values in
 contracts — read them there.
 
-**Backward compatibility**: Old model-specific names (qwen3.6-27B-code, qwen3.5-9b-it) still work
+**No backward compatibility**: Old model-specific names (qwen3.6-27B-code, qwen3.5-9b-it) are retired as of 2026-09-12
 but are deprecated for agent configs. Only the stable aliases survive model swaps. `gemma-4-12b`
 is retired and returns 400 `Invalid model name`.
 
@@ -207,7 +207,7 @@ If no SSH access, send Zulip DM via abiba-bot with vault update instructions.
 - **Router startup race**: Compose router.py doesn't call load_roster(). Reload thread sleeps 30s first.
   Fix: trigger roster reload via SSH after restart, or rebuild image with startup load_roster().
 - **LiteLLM /metrics**: Requires auth. Prometheus uses `/health/liveliness` as workaround.
-- **Strix Halo GPU**: Vulkan is the working backend (ROCm/HIP path abandoned — HSA runtime blocked on Debian 13). Build at `/root/llama.cpp/build-vk/`, commit `4fc4ec5` (2026-07-01), ggml 0.15.3 shared-lib arch. Mesa RADV 25.0.7, KHR_coopmat fast path active. ~70 tok/s gen, 532 tok/s prompt. Service: `strix-server.service` on port 8080, model: `qwen3.6-35B-udq4`, alias `strix-moe`, 128K context, flash-attn + q4 KV, multimodal (mmproj loaded).
+- **Strix Halo GPU**: Vulkan is the working backend (ROCm/HIP path abandoned — HSA runtime blocked on Debian 13). Build at `/root/llama.cpp/build-vk/`, commit `4fc4ec5` (2026-07-01), ggml 0.15.3 shared-lib arch. Mesa RADV 25.0.7, KHR_coopmat fast path active. ~70 tok/s gen, 532 tok/s prompt. Service: `strix-server.service` on port 8080, model: `Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf`, alias `strix-moe`, 256K context (n_ctx 262144), --parallel 2 --kv-unified, flash-attn + q4 KV, multimodal (mmproj loaded).
 - **Port conflict detection (2026-07-05)**: All 3 GPU wrappers now detect ghost processes squatting port 8080 before starting. `.8` and `.110` use inline pre-start check in `llama-wrapper.sh`; `.15` uses `/usr/local/bin/port-cleanup.sh` ExecStartPre. Replaces the blanket `pkill -9 -x llama-server` on .15 which would kill ALL llama-server instances regardless of port. Ghost detection was the root cause of .8 crash-looping for 27+ restarts (stale pid 25836 squatting 8080 after OOM kill).
 - **Strix Halo thermal safeguard (2026-07-02)**: `strix-server.service` has `-n 8192` (hard generation cap per request). Without it, `--predict` defaults to -1 (infinity) — a runaway request from .123 (old Mumuni CT114 — now inside Abiba CT100 at .24) decoded 39,868 tokens over 24 min, pushing Tctl to 98°C (crit 89.8°C) and throttling 70→29 t/s. The cap bounds worst-case generation to ~5 min. Do NOT remove `-n` without a replacement ceiling. Sustained load hits ~84°C even at 92s; the APU is fanless/low-flow. Clients MUST also set `max_tokens`.
 - **Port 8080 firewall**: amdpve iptables restricts 8080 to 192.168.68.116 (LiteLLM/router host) only. All inbound connections are from .116 (LiteLLM proxied via nginx). Localhost curls hang (SYN dropped). Always test from .116.
@@ -223,10 +223,10 @@ If no SSH access, send Zulip DM via abiba-bot with vault update instructions.
 | GPU | Model | Gen tok/s | Prompt tok/s | Baseline | Context |
 |-----|-------|-----------|--------------|----------|---------|
 | RTX 3090 (.8) | Qwen3.8-27B-Uncensored-Q4_K_M | **TBD** | — | — | **128K** |
-| Strix Halo (.15) | qwen3.6-35B-udq4 | **65** | 140 | — | **128K** |
+| Strix Halo (.15) | Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf (strix-moe) | **65** | 140 | — | **256K** |
 
-Benchmarks from 2026-07-17. Strix Halo model: qwen3.6-35B-udq4. RTX 5070 MTP provides 2.7x speedup over pre-upgrade 70 tok/s.
-All 3 GPUs now at 128K context (2026-07-17, reduced from 256K for stability).
+Benchmarks from 2026-07-17. Strix Halo model: Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf (alias strix-moe), n_ctx 262144, --parallel 2 --kv-unified. RTX 5070 MTP provides 2.7x speedup over pre-upgrade 70 tok/s.
+GPU contexts: RTX 3090 (.8) and RTX 5070 (.110) at 128K; Strix Halo (.15) at 256K (2026-09-12).
 
 Benchmarks run through LiteLLM proxy (192.168.68.116:4000) every 5 minutes.
 Degradation alerts fire at 30% (warning) and 50% (critical) below baseline.

gpu-self-heal.prose.md

@@ -48,7 +48,7 @@ depends_on:
 
 | Alias | GPU | Host | Model | VRAM | Ctx | tok/s | Role |
 |-------|-----|------|-------|------|-----|-------|------|
-| `strix-moe` | Strix Halo 64GB | ct15 (.15:8080) | qwen3.6-35B-udq4 | ~10/64GB (16%) | 128K | 62.9 | Compression, summarization, long docs |
+| `strix-moe` | Strix Halo 64GB | ct15 (.15:8080) | Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf | ~10/64GB (16%) | 256K | 62.9 | Compression, summarization, long docs |
 
 Key notes:
 - All models use direct GPU routing via LiteLLM (`api_key: not-needed`). Router (port 9000) was decommissioned 2026-09-11 and is NOT in the inference path.
@@ -145,7 +145,7 @@ Key notes:
 - **Detect**: Benchmark tok/s vs baseline for each GPU at current context (all 128K)
   - RTX 3090 (128K ctx, ThinkingCap): baseline 74.8 tok/s — currently at 74.9 (100%)
   - RTX 5070 (128K ctx, HauhauCS QAT): baseline 165.2 tok/s — currently at 169.6 (103%)
-  - Strix Halo (128K ctx, qwen3.6-35B-udq4): baseline 70.5 tok/s — currently at 62.9 (89%)
+  - Strix Halo (256K ctx, Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf): baseline 70.5 tok/s — currently at 62.9 (89%)
 - **Fix**:
   - If tok/s > baseline → context has headroom, consider increasing
   - If tok/s < 90% baseline → reduce context by 25% and retest

hermes-agent-baseline.prose.md

@@ -200,8 +200,7 @@ the registry before applying. Key is injected via `infisical run --` wrapper at
         { "id": "syslog-auto" },
         { "id": "strix-moe" },
         { "id": "gpu-dense" },
-        { "id": "gpu-vision" },
-        { "id": "qwen3.6-27B-code" }
+        { "id": "gpu-vision" }
       ]
     }
   }

hermes-config-template.prose.md

@@ -92,7 +92,7 @@ work immediately after restart.
 ```yaml
 # ─── Model Selection ───
 model:
-  default: <agent_model>          # e.g., strix-moe, qwen3.6-27B-code, syslog-auto
+  default: <agent_model>          # e.g., strix-moe, gpu-dense, syslog-auto
   provider: harness
   base_url: http://192.168.68.116/litellm/v1   # Rule 5 (2026-08-09): canonical authenticated path; /v1 also OK
   api_key_env: LITELLM_API_KEY    # Injected via infisical run -- wrapper
@@ -149,7 +149,7 @@ compression:
 # Do NOT use syslog-auto for auxiliary tasks — it routes to the primary GPU.
 # gpu-vision = RTX 5070 (12B), freeing the Strix Halo for agent reasoning.
 # Heavy aux (delegation, x_search) use gpu-dense (RTX 3090) instead.
-# NEVER use raw model names (e.g. qwen3.6-27B-code, qwen3.6-35B-udq4; gemma-4-12b is retired
+# NEVER use retired model names (qwen3.6-27B-code, qwen3.6-35B-udq4; gemma-4-12b is retired
 # and no longer resolves) in agent configs — use the stable aliases so model swaps don't break agents.
 auxiliary:
   vision:
@@ -173,10 +173,10 @@ auxiliary:
     timeout: 300                 # gpu-fleet: 300s for large-history summarization (was 60)
 
 # ─── Delegation / Heavy Aux (use gpu-dense = RTX 3090) ───
-# delegation.model and x_search.model use gpu-dense (NOT raw qwen3.6-27B-code).
+# delegation.model and x_search.model use gpu-dense (NOT retired raw name).
 
 delegation:
-  model: gpu-dense              # stable alias for RTX 3090 (was raw qwen3.6-27B-code)
+  model: gpu-dense              # stable alias for RTX 3090
   provider: harness
   base_url: http://192.168.68.116/litellm/v1   # Rule 5 (2026-08-09): canonical authenticated; /v1 also OK
   api_key_env: LITELLM_API_KEY
@@ -273,7 +273,7 @@ The following MUST be identical across ALL profiles:
 - The `compression: max_context_window: 131072` MUST match actual GPU capacity (128K)
 
 ### Rule 8: GPU Workload Distribution (UPDATED 2026-07-16)
-- **RTX 3090 (24GB, 128K ctx, qwen3.6-27B-code)**: Heavy reasoning, code gen, long conversations
+- **RTX 3090 (24GB, 128K ctx, gpu-dense)**: Heavy reasoning, code gen, long conversations
 - **RTX 5070 (12GB, 128K ctx, gpu-vision)**: Vision, web search, quick tasks, web_extract (IQ4_NL+MTP, ~65% VRAM at 128K)
 - **Strix Halo (64GB, 128K ctx, syslog-auto)**: Context compression, summarization, long docs
 - Agent profiles MUST route auxiliary tasks to the correct GPU:

hermes-key-enforcement.prose.md

@@ -182,7 +182,7 @@ The agent picks up the new key via `infisical run --` at gateway startup.
 # re-verify before applying.
 litellm_settings:
   default_key_generate_params:
-    models: ["syslog-auto", "qwen3.6-27B-code", "gpu-vision"]
+    models: ["syslog-auto", "gpu-dense", "gpu-vision", "strix-moe"]
     duration: null        # ← permanent
     max_budget: 100
     metadata:

inference-optimization.prose.md

@@ -101,5 +101,5 @@ call enable-prompt-caching
 
 call verify-latency
   host: 192.168.68.116
-  models: [syslog-auto, qwen3.6-27B-code, gpu-vision]
+  models: [syslog-auto, gpu-dense, gpu-vision, strix-moe]
 ```

litellm-client-timeouts.prose.md

@@ -26,9 +26,8 @@ description: >
 | Model | avg latency | avg TTFT | p-profile (24h) |
 |---|---|---|---|
 | syslog-auto | 28.8s | 25.4s | 68 calls took 30-120s; tail to ~300s under load |
-| qwen3.6-27B-code | 23.0s | — | same backend class as syslog-auto |
 | strix-moe | 7.5s | — | Strix Halo, healthy |
-| gemma-4-12b (retired 2026-09-12; RTX 5070 now `gpu-vision`) | 2.6s | — | RTX 5070, healthy |
+| gpu-vision (retired gemma-4-12b, RTX 5070) | 2.6s | — | RTX 5070, healthy |
 
 Sep 6 incident timeline: failures 04:00-07:00 EDT (0% GPU util = wedged
 backend), full recovery 07:00-08:00 with ZERO client failures once requests
@@ -56,9 +55,8 @@ proxy queuing.
 - vision: 60s (keep), web_extract: 30s (keep) — the 2.6s average was measured on `gemma-4-12b` (retired 2026-09-12); the live RTX 5070 alias is `gpu-vision`.
 - compression: 300s (keep — this was already raised from 60 per gpu-fleet).
 - **gpu-dense delegation/x_search: set timeout >= 120s.** The RTX 3090
-  (qwen3.6-27B-code backend, 23.0s avg) is the same speed class as
-  syslog-auto; delegation defaults that assume fast responses will 408 the
-  same way.
+  (gpu-dense backend) is the same speed class as syslog-auto; delegation
+  defaults that assume fast responses will 408 the same way.
 
 ### 3. Retry policy — backoff, not repetition
 

litellm-health.prose.md

@@ -144,7 +144,7 @@ contracts — read them there. Do not re-add retired names (`gemma-4-12b`, `gpu-
 7. **Check model inference via LiteLLM** — Test one model on each GPU host. The health
    check runs on the **backend edge**, not the public edge, so these paths carry the
    `/litellm/` prefix:
-   - POST http://{{backend_host}}/litellm/v1/chat/completions model=qwen3.6-27B-code → expect 200 (RTX 3090, .8)
+   - POST http://{{backend_host}}/litellm/v1/chat/completions model=gpu-vision → expect 200 (RTX 5070, .110)
    - POST http://{{backend_host}}/litellm/v1/chat/completions model=gpu-vision → expect 200 (RTX 5070, .110)
    - POST http://{{backend_host}}/litellm/v1/chat/completions model=strix-moe → expect 200 (Strix Halo, .15)
    - Auth uses the dedicated `monitor` agent key, read on CT 116 from

pm2-self-heal.prose.md

@@ -9,7 +9,6 @@ description: >
 - abiba-telegram: { status: "online", uptime: string, restarts: number }
 - abiba-zulip: { status: "online", uptime: string, restarts: number }
 - gitea-runner: { status: "online", uptime: string, restarts: number }
-- spoton-service: { status: "online", uptime: string, restarts: number }
 - zulip-watchdog: { status: "online", uptime: string, restarts: number }
 - last_check: timestamp
 

tests/test_audit_hermes_config_alias.py

@@ -132,20 +132,20 @@ def test_retired_alias_in_custom_providers_is_rejected(tmp_path):
     assert "RESULT: FAIL" in out
 
 
-def test_raw_but_live_alias_warns_but_passes(tmp_path):
-    """Raw-but-live names resolve (200), so they warn only; failing them rejects valid configs."""
+def test_retired_raw_name_fails(tmp_path):
+    """Retired raw names no longer resolve (400), so they fail; the 2026-09-12 registry change moved qwen3.6-27B-code from raw-but-live to non-resolving."""
     code, out = _run_config(
         tmp_path,
-        "raw-qwen.yaml",
+        "retired-qwen.yaml",
         BASE.format(alias="gpu-vision").replace(
             "delegation:\n  provider: harness",
             "delegation:\n  provider: harness\n  model: qwen3.6-27B-code",
         ),
     )
-    assert code == 0, out
-    assert "delegation.model = 'qwen3.6-27B-code' is a raw-but-live model name" in out
-    assert "prefer the stable alias gpu-dense" in out
-    assert "RESULT: PASS" in out
+    assert code != 0, out
+    assert "delegation.model = 'qwen3.6-27B-code' is retired and no longer resolves" in out
+    assert "use gpu-dense" in out
+    assert "RESULT: FAIL" in out
 
 
 def test_retired_alias_in_fallback_providers_is_rejected(tmp_path):