fix(cli): honour NEMOCLAW_LOCAL_INFERENCE_TIMEOUT for compatible-endpoint (#2403)

[jason-ma-nv]

Summary

compatible-endpoint skipped the --timeout flag on openshell inference set, leaving the 60 s default active even when NEMOCLAW_LOCAL_INFERENCE_TIMEOUT was exported. ollama-local and vllm-local already pass the flag; this brings compatible-endpoint into parity so reasoning-model operators on external LAN endpoints are not silently capped at 60 seconds.

Related Issue

Fixes #2403

Changes

• src/lib/onboard.ts — add --timeout to inference set args when provider is compatible-endpoint, using the same LOCAL_INFERENCE_TIMEOUT_SECS constant already applied to ollama-local and vllm-local
• test/onboard-selection.test.ts — new test that spawns a fake openshell binary recording inference set args, sets NEMOCLAW_LOCAL_INFERENCE_TIMEOUT=600, calls setupInference directly, and asserts --timeout 600 is present

Type of Change

• Code change (feature, bug fix, or refactor)
• Code change with doc updates
• Doc only (prose changes, no code sample modifications)
• Doc only (includes code sample changes)

Verification

• npx prek run --all-files passes
• npm test passes
• Tests added or updated for new or changed behavior
• No secrets, API keys, or credentials committed
• Docs updated for user-facing behavior changes
• make docs builds without warnings (doc changes only)
• Doc pages follow the style guide (doc changes only)
• New doc pages includes SPDX header and frontmatter (new pages only)

AI Disclosure

• AI-assisted — tool: Claude Code

---

Signed-off-by: Jason Ma jama@nvidia.com

Summary by CodeRabbit

• New Features

  • Added configurable timeout support for local inference setup, applied for the compatible endpoint flow.

• Tests

  • Added a regression test to verify the configured timeout is forwarded to the inference initialization command.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds plumbing so NEMOCLAW_LOCAL_INFERENCE_TIMEOUT (via LOCAL_INFERENCE_TIMEOUT_SECS) is forwarded as --timeout when setupInference configures the compatible-endpoint provider; includes a regression test verifying the value is passed into the openshell inference set invocation.

Changes

Cohort / File(s)	Summary
Timeout configuration for compatible-endpoint src/lib/onboard.ts	Conditionalizes openshell inference set argument construction to include --timeout <LOCAL_INFERENCE_TIMEOUT_SECS> for the compatible-endpoint provider.
Regression test for timeout propagation test/onboard-selection.test.ts	Adds a test that sets NEMOCLAW_LOCAL_INFERENCE_TIMEOUT, injects a fake openshell binary to capture inference set args, and asserts --timeout 600 is present for the compatible-endpoint flow.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

Poem

🐇 I hopped through code with whiskers twitching bright,
A timeout carried home from dark to light,
Compatible-endpoint now holds the key,
Models can think a while — set them free! 🥕

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes the main change: fixing the compatible-endpoint provider to honour the NEMOCLAW_LOCAL_INFERENCE_TIMEOUT environment variable, which directly addresses issue #2403.
Linked Issues check	✅ Passed	The PR fully addresses the coding requirements from issue #2403: plumbing NEMOCLAW_LOCAL_INFERENCE_TIMEOUT into compatible-endpoint provider so it passes --timeout to openshell inference set, with corresponding regression test added.
Out of Scope Changes check	✅ Passed	All changes are in-scope: src/lib/onboard.ts adds timeout flag for compatible-endpoint provider, and test/onboard-selection.test.ts adds a regression test validating the fix. Both directly address the linked issue requirement.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/compatible-endpoint-timeout-2403

---

_{Review rate limit: 9/10 reviews remaining, refill in 6 minutes.}

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[wscurran]

✨ Thanks for submitting this PR that fixes a bug and improves the inference experience with the compatible-endpoint provider by honoring the NEMOCLAW_LOCAL_INFERENCE_TIMEOUT environment variable. This change brings the compatible-endpoint into parity with ollama-local and vllm-local, ensuring that reasoning-model operators on external LAN endpoints are not silently capped at 60 seconds.

---

Related open issues:

• #2403 compatible-endpoint provider does not honour NEMOCLAW_LOCAL_INFERENCE_TIMEOUT (vllm-local and ollama-local do); 60s default leaks through to reasoning-model streams

[ericksoa]

Looks good. This is narrowly scoped to , reuses the existing local inference timeout setting, and includes a regression test that verifies receives the expected argument.

[coderabbitai]

🧹 Nitpick comments (1)

test/onboard-selection.test.ts (1)

3456-3463: Tighten timeout assertion to validate flag/value pairing.

Current checks can pass if "600" appears elsewhere in args. Prefer asserting "600" is the immediate value after "--timeout".

Proposed assertion hardening

-    assert.ok(
-      state.inferenceSetArgs.includes("--timeout"),
-      `Expected --timeout in inference set args, got: ${JSON.stringify(state.inferenceSetArgs)}`,
-    );
-    assert.ok(
-      state.inferenceSetArgs.includes("600"),
-      `Expected 600 in inference set args, got: ${JSON.stringify(state.inferenceSetArgs)}`,
-    );
+    const timeoutIdx = state.inferenceSetArgs.indexOf("--timeout");
+    assert.notEqual(
+      timeoutIdx,
+      -1,
+      `Expected --timeout in inference set args, got: ${JSON.stringify(state.inferenceSetArgs)}`,
+    );
+    assert.equal(
+      state.inferenceSetArgs[timeoutIdx + 1],
+      "600",
+      `Expected --timeout value 600, got: ${JSON.stringify(state.inferenceSetArgs)}`,
+    );

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@test/onboard-selection.test.ts` around lines 3456 - 3463, The test currently
just asserts that "--timeout" and "600" exist anywhere in
state.inferenceSetArgs; change it to find the index of "--timeout" in
state.inferenceSetArgs (using state.inferenceSetArgs.indexOf("--timeout")),
assert the index is >= 0 and index + 1 is within bounds, then assert that
state.inferenceSetArgs[index + 1] === "600" so the flag and value are paired.
This tightens the assertion in the failing test that uses
state.inferenceSetArgs.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@test/onboard-selection.test.ts`:
- Around line 3456-3463: The test currently just asserts that "--timeout" and
"600" exist anywhere in state.inferenceSetArgs; change it to find the index of
"--timeout" in state.inferenceSetArgs (using
state.inferenceSetArgs.indexOf("--timeout")), assert the index is >= 0 and index
+ 1 is within bounds, then assert that state.inferenceSetArgs[index + 1] ===
"600" so the flag and value are paired. This tightens the assertion in the
failing test that uses state.inferenceSetArgs.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: d431ed71-5511-4b15-a110-e6a47baa1ddc

📥 Commits

Reviewing files that changed from the base of the PR and between 9e2b975 and db7e8eb.

📒 Files selected for processing (2)

• src/lib/onboard.ts
• test/onboard-selection.test.ts

🚧 Files skipped from review as they are similar to previous changes (1)

• src/lib/onboard.ts