A self-improving RLM agent for coding workflows and long-running autonomous tasks.
Rust
fix ci after model catalog refresh (#324)
kt6373d171fa2d
diff --git a/packages/ai/test/prime-inference-models.test.ts b/packages/ai/test/prime-inference-models.test.tsindex 27acc1aaf8..cb5e0bc9b0 100644--- a/packages/ai/test/prime-inference-models.test.ts+++ b/packages/ai/test/prime-inference-models.test.ts@@ -30,7 +30,6 @@ describe("Prime Inference models", () => {"nvidia/nemotron-3-super-120b-a12b","openai/gpt-5.4","openai/gpt-5.5",- "prime-intellect/intellect-3","qwen/qwen3-coder-next","qwen/qwen3-vl-235b-a22b-thinking","x-ai/grok-4.20",diff --git a/packages/coding-agent/examples/sdk/12-full-control.ts b/packages/coding-agent/examples/sdk/12-full-control.tsindex 2e6fd63662..1c1cdbca43 100644--- a/packages/coding-agent/examples/sdk/12-full-control.ts+++ b/packages/coding-agent/examples/sdk/12-full-control.ts@@ -26,7 +26,7 @@ if (process.env.MY_ANTHROPIC_KEY) {// Model registry with no custom models.jsonconst modelRegistry = ModelRegistry.inMemory(authStorage);-const model = getModel("anthropic", "claude-sonnet-4-20250514");+const model = getModel("anthropic", "claude-sonnet-5");if (!model) throw new Error("Model not found");// In-memory settings with overridesdiff --git a/packages/coding-agent/test/agent-session-auto-compaction-queue.test.ts b/packages/coding-agent/test/agent-session-auto-compaction-queue.test.tsindex f53431700c..c23ff38c20 100644--- a/packages/coding-agent/test/agent-session-auto-compaction-queue.test.ts+++ b/packages/coding-agent/test/agent-session-auto-compaction-queue.test.ts@@ -237,6 +237,7 @@ describe("AgentSession auto-compaction queue resume", () => {it("should trigger threshold compaction for error messages using last successful usage", async () => {const model = session.model!;+ const highUsageTokens = model.contextWindow;// A successful assistant message with high token usage (near context limit)const successfulAssistant: AssistantMessage = {@@ -246,11 +247,11 @@ describe("AgentSession auto-compaction queue resume", () => {provider: model.provider,model: model.id,usage: {- input: 180_000,- output: 10_000,+ input: highUsageTokens,+ output: 0,cacheRead: 0,cacheWrite: 0,- totalTokens: 190_000,+ totalTokens: highUsageTokens,cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },},stopReason: "stop",diff --git a/packages/coding-agent/test/kernel-state-roundtrip.test.ts b/packages/coding-agent/test/kernel-state-roundtrip.test.tsindex d7164571a2..e79d88d927 100644--- a/packages/coding-agent/test/kernel-state-roundtrip.test.ts+++ b/packages/coding-agent/test/kernel-state-roundtrip.test.ts@@ -51,13 +51,12 @@ describeIfKernel("kernel state snapshot round-trip (real kernel)", () => {await writer.execute("x = 42");await writer.execute("df = [1, 2, 3]");await writer.execute("def double(n):\n return n * 2");- await writer.execute("import socket\nsock = socket.socket()");+ await writer.execute("gen = (n for n in range(3))");const snap = await writer.snapshotState();expect(snap).not.toBeNull();expect(snap?.saved).toEqual(expect.arrayContaining(["x", "df", "double"]));- // A live socket cannot be pickled and must be reported, not silently dropped.- expect(snap?.skipped.map((s) => s.name)).toContain("sock");+ expect(snap?.skipped.map((s) => s.name)).toContain("gen");expect(existsSync(snapshotPath)).toBe(true);expect(existsSync(manifestPath)).toBe(true);} finally {