feat(onboard): add MiniMax provider

[octo-patch]

Reason: add target provider/model to existing provider registry

Summary

• Register MiniMax as a first-class OpenAI-compatible onboarding provider with MiniMax-M3 as the default model.
• Add MiniMax to the provider selection menu and non-interactive provider keys.
• Route minimax-api selections through the sandbox inference endpoint with MINIMAX_API_KEY.

Test plan

• git diff --check
• secret scan over git diff HEAD
• npx -p node@22.16.0 -p npm@10.8.2 npm --prefix /root/octopatch-4/work/repo-recvoNmbfQOgyd run typecheck
• npx -p node@22.16.0 bash -lc 'cd /root/octopatch-4/work/repo-recvoNmbfQOgyd && node node_modules/vitest/vitest.mjs --config vitest.config.ts run --project cli src/lib/onboard/providers.test.ts src/lib/onboard/provider-menu.test.ts src/lib/inference/config.test.ts'
• npx -p node@22.16.0 bash -lc 'cd /root/octopatch-4/work/repo-recvoNmbfQOgyd/nemoclaw && node node_modules/typescript/bin/tsc --noEmit --incremental'

🤖 Generated with Claude Code

Summary by CodeRabbit

• New Features

  • Added MiniMax as a supported provider in onboarding and inference selection.
  • MiniMax can now be chosen from provider menus and works with the same remote-provider flow as other supported options.

• Bug Fixes

  • Provider selection now recognizes MiniMax correctly and falls back to the expected default model when one isn’t specified.
  • Updated validation and provider ordering so MiniMax appears in the right places across setup and configuration screens.

[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]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: ab9e480a-b64c-4623-9feb-018f2c603d83

📥 Commits

Reviewing files that changed from the base of the PR and between 7215882 and 01c6c52.

📒 Files selected for processing (6)

• src/lib/inference/config.test.ts
• src/lib/inference/config.ts
• src/lib/onboard/provider-menu.test.ts
• src/lib/onboard/provider-menu.ts
• src/lib/onboard/providers.test.ts
• src/lib/onboard/providers.ts

---

📝 Walkthrough

Walkthrough

Adds MiniMax as a new supported inference/onboarding provider: a minimax-api case is added to inference config selection, a minimax entry is registered in REMOTE_PROVIDER_CONFIG and the non-interactive provider allowlist, and a minimax option is added to the provider menu, with corresponding tests updated.

Changes

MiniMax Provider Integration

Layer / File(s)	Summary
MiniMax provider registration src/lib/onboard/providers.ts, src/lib/onboard/providers.test.ts	Adds MINIMAX_ENDPOINT_URL, includes minimax in NON_INTERACTIVE_PROVIDER_KEYS/valid values, adds a REMOTE_PROVIDER_CONFIG.minimax entry with provider metadata and credential env, and tests these registrations.
Inference selection config src/lib/inference/config.ts, src/lib/inference/config.test.ts	Adds a minimax-api case to getProviderSelectionConfig returning routed endpoint, MINIMAX_API_KEY, MiniMax label, and default model MiniMax-M3; tests verify shape, approved provider set, and default fallback.
Provider menu entry src/lib/onboard/provider-menu.ts, src/lib/onboard/provider-menu.test.ts	Adds { key: "minimax", label: "MiniMax" } to BASE_REMOTE_PROVIDER_OPTIONS; tests update expected option key ordering to include minimax.

Estimated code review effort: 2 (Simple) | ~10 minutes

Possibly related PRs

• NVIDIA/NemoClaw#5165: Both PRs modify inference provider menu construction in src/lib/onboard/provider-menu.ts.

Suggested labels: area: cli, area: onboarding, v0.0.64

🚥 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 is concise and accurately summarizes the main change: adding MiniMax as a provider.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[cv]

Thanks for the MiniMax contribution. Before a maintainer can vet this copy PR or run trusted CI, the contributor-compliance gates need to be fixed:

• The current commit 01c6c5206840ecbb71b57b7ba1acb3294cc5b3ac is unsigned and GitHub reports it as Unverified.
• The PR description does not include the required contributor Signed-off-by: Name <email> DCO declaration.

Please replace the branch history with GitHub-Verified commit(s) using the repository contributor workflow and add the matching Signed-off-by: declaration to the PR body. If this published branch cannot be safely rewritten, open a fresh branch and PR with compliant history. Once both are present, we can proceed with runner vetting and the substantive provider review.

[cjagwani]

Follow-up for the compliant-history refresh: the current exact head also has substantive CI gaps that should be fixed in the same contributor-owned replacement.

• cli-parity: add minimax to the provider usage in install.sh, scripts/install.sh, and the NEMOCLAW_PROVIDER row in docs/reference/commands.mdx.
• build-typecheck / cli-test-shards (5): MiniMax shifts numbered provider choices, so the compatible-endpoint and Anthropic retry fixtures now select the wrong provider, prompt for unrelated credentials, and eventually OOM. Please update the affected menu fixtures—prefer label-based selection where the harness supports it—and audit the other numbered custom/Ollama/NIM/vLLM cases.
• Credential lifecycle: add or justify MINIMAX_API_KEY in the canonical known-credential/reset path and cover non-interactive MiniMax onboarding/provider-key resolution.

The existing compliance request remains the first hard gate: the PR body needs the contributor's Signed-off-by: declaration and the replacement commit history must appear Verified in GitHub. Maintainers will not rewrite that history on the contributor's behalf.

[github-actions]

E2E Advisor Recommendation

Required E2E: inference-routing, network-policy, onboard-repair, onboard-resume, cloud-onboard
Optional E2E: None

Dispatch hint: inference-routing,network-policy,onboard-repair,onboard-resume,cloud-onboard

Workflow run

Full advisor summary

E2E Recommendation Advisor

Base: target/main
Head: HEAD
Confidence: high

Required E2E

• inference-routing: Required by the deterministic risk plan for inference-policy changes. The new MiniMax provider config must still route through the managed sandbox inference endpoint and preserve real host-to-sandbox inference behavior.
• network-policy: Required by the deterministic risk plan for inference-policy changes. Provider route changes must continue to agree with the live network policy boundary: intended inference routes are allowed and unintended egress is denied.
• onboard-repair: Required by the deterministic risk plan for lifecycle-state changes. Onboarding provider metadata changes can affect repaired onboarding state, gateway provider metadata, and sandbox convergence after partial failure.
• onboard-resume: Required by the deterministic risk plan for lifecycle-state changes. Provider menu and registry changes can affect interrupted onboarding resume with cached setup and provider state restoration.
• cloud-onboard: The PR changes real onboarding provider choices, non-interactive provider validation, credential metadata, and provider registration. Run the hosted onboarding flow to catch regressions in the end-to-end user onboarding path and provider prompt/menu behavior.

Optional E2E

• None.

New E2E recommendations

• minimax-provider-smoke (medium): Existing PR-safe inference-routing coverage exercises routed/fake-compatible paths, while credential-backed provider smokes are intentionally kept out of the PR lane. This PR adds MiniMax as a first-class hosted provider, but there appears to be no MiniMax-specific trusted live smoke validating real credentials, endpoint compatibility, and sandbox inference response.
  • Suggested test: Add a trusted-main credential-backed MiniMax onboarding and inference-routing smoke that registers minimax-api with MINIMAX_API_KEY, verifies provider CRUD, confirms sandbox inference.local reaches https://api.minimax.io/v1, and checks credential isolation.

Dispatch hint

• Workflow: .github/workflows/e2e.yaml
• jobs input: inference-routing,network-policy,onboard-repair,onboard-resume,cloud-onboard

[cv]

Release sweep: deferring this PR from v0.0.81.

The MiniMax integration remains unique work, but it is not releasable in its current contributor history: the feature commit is not GitHub Verified, the PR body lacks the required Signed-off-by: declaration, and the required checks are absent. The outstanding CLI/help/docs parity, credential lifecycle coverage, onboarding fixture repair, and end-to-end route coverage also need to be included in a fresh compliant submission.

Please keep the replacement narrowly scoped and re-request targeting once those gates are satisfied.

[octo-patch]

Replaced by #6761 from a clean branch based on current main. Closing this PR in favor of the replacement.