fix(coding-agent): include llama.cpp in append-only allowlist

ModelRegistry registers a built-in llama.cpp discovery as `provider:
"llama.cpp"` (model-registry.ts:1063), so a reverse-proxied or
public-DNS llama.cpp endpoint never trips the loopback heuristic and
was still falling through to no append-only context.

Add "llama.cpp" to LOCAL_INFERENCE_PROVIDERS, refresh the docstring
on `hasLocalLoopbackBaseUrl` (it covers user-defined providers; built-
in local ids are caught by the allowlist), and extend the test to
exercise a public-host llama.cpp baseUrl so the allowlist path is
covered independently of the loopback heuristic.

Refs #3033

packages/coding-agent/CHANGELOG.md

@@ -4,7 +4,7 @@
 
 ### Fixed
 
-- Fixed Ollama (and LM Studio / loopback llama.cpp / vLLM servers) reprocessing the full prompt on every turn because `provider.appendOnlyContext: auto` only recognized DeepSeek and Xiaomi as prefix-cache providers. The auto-detect now enables append-only mode for `ollama`, `ollama-cloud`, `lm-studio`, and any baseUrl resolving to a loopback/RFC1918/`.local` host, so the system prompt + tool catalogue + prior-turn message bytes stay byte-stable across turns and llama.cpp's KV-cache prefix reuse can hit ([#3033](https://github.com/can1357/oh-my-pi/issues/3033)).
+- Fixed Ollama, LM Studio, and llama.cpp (plus loopback vLLM / sglang servers) reprocessing the full prompt on every turn because `provider.appendOnlyContext: auto` only recognized DeepSeek and Xiaomi as prefix-cache providers. The auto-detect now enables append-only mode for `ollama`, `ollama-cloud`, `lm-studio`, `llama.cpp`, and any baseUrl resolving to a loopback/RFC1918/`.local` host, so the system prompt + tool catalogue + prior-turn message bytes stay byte-stable across turns and llama.cpp's KV-cache prefix reuse can hit ([#3033](https://github.com/can1357/oh-my-pi/issues/3033)).
 
 ## [16.1.1] - 2026-06-19
 

packages/coding-agent/src/config/append-only-context-mode.ts

@@ -16,12 +16,14 @@ export interface AppendOnlyContextModel {
  * tool catalogue, and message log all flow through fresh allocations every
  * step (see `agent-loop.ts` `streamAssistantResponse` fallback path).
  */
-const LOCAL_INFERENCE_PROVIDERS = new Set(["ollama", "ollama-cloud", "lm-studio"]);
+const LOCAL_INFERENCE_PROVIDERS = new Set(["ollama", "ollama-cloud", "lm-studio", "llama.cpp"]);
 
-/** True when `baseUrl` resolves to a loopback or RFC1918 host — the heuristic
- * for user-defined providers (llama.cpp / vLLM via `models.yaml`) that omp
- * has no provider id for. Substring match on parsed hostname only; ports,
- * paths, and unparseable URLs return false.
+/** True when `baseUrl` resolves to a loopback or RFC1918 host — covers
+ * llama.cpp/vLLM/sglang servers registered under a user-defined provider id
+ * via `models.yaml`. Built-in local provider ids (`ollama`, `lm-studio`,
+ * `llama.cpp`) are already handled by `LOCAL_INFERENCE_PROVIDERS`.
+ * Substring match on the parsed hostname only; ports, paths, and unparseable
+ * URLs return false.
  */
 function hasLocalLoopbackBaseUrl(baseUrl: string | undefined): boolean {
 	if (!baseUrl) return false;

packages/coding-agent/test/append-only-context-mode.test.ts

@@ -55,6 +55,14 @@ describe("shouldEnableAppendOnlyContext", () => {
 		expect(
 			shouldEnableAppendOnlyContext("auto", { provider: "lm-studio", baseUrl: "http://127.0.0.1:1234/v1" }),
 		).toBe(true);
+		// `llama.cpp` is a built-in provider id (ModelRegistry registers it for keyless local discovery);
+		// the allowlist must catch it even when the user reverse-proxies the server through a public host.
+		expect(
+			shouldEnableAppendOnlyContext("auto", { provider: "llama.cpp", baseUrl: "https://llamacpp.example.com/v1" }),
+		).toBe(true);
+		expect(shouldEnableAppendOnlyContext("auto", { provider: "llama.cpp", baseUrl: "http://127.0.0.1:8080" })).toBe(
+			true,
+		);
 	});
 
 	test("auto enables for loopback and private baseUrls (user-defined llama.cpp/vLLM)", () => {