feat: add minimax-m2.7 to model registry

AGENTS.md

@@ -60,10 +60,11 @@ These decisions are part of the repo contract. Changing them is not a small refa
 - local scope must protect `.claude/settings.local.json` from git before secrets are written
 - settings files must be written with owner-only permissions
 
-Current honest limitation:
+Current curated public Claude Code model registry:
 
-- the curated registry currently contains exactly one supported model:
-  `qwen3-235b` -> `qwen/qwen3-235b-a22b-instruct-2507-fp8`
+- `qwen3-235b` -> `qwen/qwen3-235b-a22b-instruct-2507-fp8`
+- `kimi-k2.6` -> `moonshotai/Kimi-K2.6` (default)
+- `minimax-m2.7` -> `minimaxai/minimax-m2.7`
 
 ## What the Repo Does and Does Not Do
 

CHANGELOG.md

@@ -2,6 +2,7 @@
 
 ## Unreleased
 
+- Added curated model support for `minimax-m2.7` -> `minimaxai/minimax-m2.7` in the installer model picker and docs.
 - Added curated model support for `kimi-k2.6` -> `moonshotai/Kimi-K2.6` in the installer model picker and docs.
 - Made local alias package bootstrap publish without npm provenance because provenance requires a supported CI/OIDC provider.
 - Made alias package publishing skip cleanly until `@gonkagate/claude-code-setup` is bootstrapped on npm, with an explicit override for the first package publish.

README.md

@@ -60,6 +60,7 @@ Current public Claude Code models in the curated registry:
 
 - `qwen3-235b` -> `qwen/qwen3-235b-a22b-instruct-2507-fp8`
 - `kimi-k2.6` -> `moonshotai/Kimi-K2.6` (default)
+- `minimax-m2.7` -> `minimaxai/minimax-m2.7`
 
 ## What It Does
 

docs/how-it-works.md

@@ -16,6 +16,7 @@ Today the curated public Claude Code model registry contains:
 
 - `qwen3-235b` -> `qwen/qwen3-235b-a22b-instruct-2507-fp8`
 - `kimi-k2.6` -> `moonshotai/Kimi-K2.6` (default)
+- `minimax-m2.7` -> `minimaxai/minimax-m2.7`
 
 Users provide only:
 

docs/troubleshooting.md

@@ -18,6 +18,7 @@ Today that curated list contains:
 
 - `qwen3-235b` -> `qwen/qwen3-235b-a22b-instruct-2507-fp8`
 - `kimi-k2.6` -> `moonshotai/Kimi-K2.6` (default)
+- `minimax-m2.7` -> `minimaxai/minimax-m2.7`
 
 If that model is unavailable, the likely cause is a backend deployment or model availability mismatch. This installer does not expose custom base URL or arbitrary custom model overrides, so the right next step is GonkaGate support or the backend troubleshooting docs in `gonka-proxy`.
 

src/constants/models.ts

@@ -20,6 +20,13 @@ const curatedModelRegistry = [
     modelId: "moonshotai/Kimi-K2.6",
     description: "GonkaGate public Claude Code model.",
     isDefault: true
+  },
+  {
+    key: "minimax-m2.7",
+    displayName: "MiniMax M2.7",
+    modelId: "minimaxai/minimax-m2.7",
+    description: "GonkaGate public Claude Code model.",
+    isDefault: false
   }
 ] as const satisfies readonly SupportedModelDefinition[];
 

test/install.test.ts

@@ -87,6 +87,12 @@ test("supported model registry includes kimi-k2.6", () => {
   assert.equal(model.modelId, "moonshotai/Kimi-K2.6");
 });
 
+test("supported model registry includes minimax-m2.7", () => {
+  const model = requireSupportedModel("minimax-m2.7");
+
+  assert.equal(model.modelId, "minimaxai/minimax-m2.7");
+});
+
 test("loadSettings rejects invalid JSON instead of overwriting it", async () => {
   const directory = await mkdtemp(path.join(tmpdir(), "gonkagate-invalid-json-"));
   const filePath = path.join(directory, "settings.json");