Skip to content

feat(cohere): add current Command models#102563

Merged
steipete merged 2 commits into
mainfrom
codex/cohere-command-refresh
Jul 9, 2026
Merged

feat(cohere): add current Command models#102563
steipete merged 2 commits into
mainfrom
codex/cohere-command-refresh

Conversation

@steipete

@steipete steipete commented Jul 9, 2026

Copy link
Copy Markdown
Contributor

Closes #102545

What Problem This Solves

OpenClaw's direct Cohere provider exposes Command A from March 2025 and North Mini Code, but omits Cohere's current Command A+, Command A Reasoning, and Command A Vision models. New Cohere users therefore get a stale onboarding default and must hand-author incomplete model metadata to use current models.

Why This Change Was Made

Add the current Command catalog with Cohere's documented modality, context, output, reasoning, and tool-use contracts; make Command A+ the onboarding default; and advertise only tool-capable current Cohere models to modern agent sweeps. The same sweep refresh selects stable Gemini 3.5 Flash, retains both Grok 4.5 and the dated Grok 4.20 reasoning release, and rejects mutable -latest-... aliases.

Command A Vision remains selectable but is explicitly non-tool-capable and excluded from tool-oriented modern sweeps.

User Impact

Users can select Command A+, Command A Reasoning, and Command A Vision without custom provider definitions. Fresh Cohere onboarding selects Command A+, thinking levels map to the Compatibility API's supported none/high values, and Vision requests do not receive unsupported agent tools.

Evidence

  • Official contracts: Command A+, Command A Reasoning, Command A Vision, and Compatibility API.
  • Blacksmith Testbox tbx_01kx2xyv9heryvxpk1re9t6hds (blacksmith-testbox, Actions run):
    • corepack pnpm test extensions/cohere/index.test.ts extensions/cohere/onboard.test.ts src/agents/model-compat.test.ts — 66 tests passed.
    • corepack pnpm test:extension cohere — 9 tests passed.
    • env OPENCLAW_CHECK_CHANGED_REMOTE_CHILD=1 OPENCLAW_CHANGED_LANES_RAW_SYNC=1 corepack pnpm check:changed — all selected typecheck, lint, docs, and guard lanes passed.
    • corepack pnpm build — passed.
  • Fresh autoreview: clean after addressing Vision sweep capability and mutable xAI alias findings.
  • GCP Crabbox cbx_5502273e1db8 (gcp, n2-standard-8, run):
    • OPENCLAW_LIVE_GATEWAY=1 OPENCLAW_LIVE_GATEWAY_PROVIDERS=cohere OPENCLAW_LIVE_GATEWAY_MODELS=cohere/command-a-plus-05-2026 OPENCLAW_LIVE_GATEWAY_MAX_MODELS=1 corepack pnpm test:live src/gateway/gateway-models.profiles.live.test.ts — 110 tests passed.
    • The real Cohere Compatibility API returned HTTP 200 for Command A+; prompt, session follow-up, read, exec + read, and image probes all passed.
    • COHERE_API_KEY was injected as one redacted, process-scoped environment variable and was not written to the checkout or logs.

AI-assisted: yes. No agent transcript attached.

@openclaw-barnacle openclaw-barnacle Bot added docs Improvements or additions to documentation gateway Gateway runtime agents Agent runtime and tooling labels Jul 9, 2026
@steipete steipete self-assigned this Jul 9, 2026
@openclaw-barnacle openclaw-barnacle Bot added size: M maintainer Maintainer-authored PR labels Jul 9, 2026
@clawsweeper

clawsweeper Bot commented Jul 9, 2026

Copy link
Copy Markdown
Contributor

Codex review: needs maintainer review before merge. Reviewed July 9, 2026, 6:06 AM ET / 10:06 UTC.

Summary
The PR adds Command A+, Command A Reasoning, and Command A Vision to the Cohere provider catalog, changes fresh Cohere onboarding to Command A+, wires Cohere into modern live sweeps, and updates related docs/tests plus Google/xAI sweep refs.

PR surface: Source +107, Tests +81, Docs +7. Total +195 across 13 files.

Reproducibility: not applicable. as a feature PR rather than a broken existing-behavior report. Source inspection shows current main lacks the requested Cohere Command entries, and the PR head adds them.

Review metrics: 3 noteworthy metrics.

  • Provider Default Surface: 1 changed. The Cohere onboarding default changes fresh setup model choice from old Command A to Command A+.
  • Provider Catalog Models: 3 added. The bundled Cohere provider gains three new static model rows that affect model selection and capability metadata.
  • Curated Live Sweep Refs: 1 replaced, 2 added. The default high-signal live matrix changes which provider/model refs operator validation lanes exercise.

Root-cause cluster
Relationship: fixed_by_candidate
Canonical: #102545
Summary: This PR is the candidate implementation for the open Cohere Command models feature request; earlier provider-catalog work only partially overlaps.

Members:

Proposal only: this assessment does not dispatch repair, suppress jobs, mutate sibling items, close, or merge anything.

Merge readiness
Overall: 🐚 platinum hermit
Proof: 🐚 platinum hermit
Patch quality: 🐚 platinum hermit
Result: ready for maintainer review.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Rank-up moves:

  • Get explicit maintainer approval for the Command A+ default/live-sweep switch after exact-head checks finish.

Risk before merge

  • [P1] Changing fresh Cohere onboarding from cohere/command-a-03-2025 to cohere/command-a-plus-05-2026 changes new-user provider model choice and should be an explicit maintainer-approved default decision.
  • [P1] The curated live sweep changes add Cohere A+ and update Google/xAI refs, so exact-head CI and maintainer approval should gate merge even though the Cohere live proof is now sufficient.
  • [P1] The PR is behind the current base; this is not a patch defect, but final merge should use the actual merge result with required checks green.

Maintainer options:

  1. Approve The Default Switch (recommended)
    A maintainer can accept the compatibility risk and merge after confirming Command A+ should be the fresh Cohere default and sweep representative.
  2. Preserve The Existing Default
    Change the PR to add the new models while leaving command-a-03-2025 as the onboarding default until the default policy is approved.
  3. Split The Sweep Refresh
    Separate the Google/xAI live-sweep edits from the Cohere catalog/default change if maintainers want a narrower merge surface.

Next step before merge

  • [P2] Manual review is needed because the PR has a protected maintainer label and changes provider defaults/live-sweep behavior; there is no narrow automated repair blocker.

Maintainer decision needed

  • Question: Should fresh Cohere onboarding and modern live sweeps move to Command A+ in this PR, or should the new models land without changing the default/sweep representative?
  • Rationale: The patch is technically coherent and live-proven, but provider defaults and model-choice policy affect new setups and operator validation behavior, which ClawSweeper should not decide automatically under the protected maintainer label.
  • Likely owner: steipete — This login has the most recent related provider-catalog history and owns the current PR/linked issue context.
  • Options:
    • Approve Command A+ Default (recommended): Merge the PR with Command A+ as the Cohere onboarding default and modern sweep model now that live A+ proof is present.
    • Keep Old Default: Land the new catalog entries but keep command-a-03-2025 as the onboarding default until default policy is separately approved.
    • Defer Direction: Keep the PR open until maintainers decide whether the Google/xAI live-sweep refresh belongs with this Cohere catalog change.

Security
Cleared: Cleared: the diff changes provider catalog metadata, docs, tests, and live-sweep selection without changing dependencies, workflows, lockfiles, permissions, or secret handling.

Review details

Best possible solution:

Land the provider-owned catalog, docs, and tests only after maintainers explicitly approve Command A+ as the fresh Cohere default and modern sweep representative.

Do we have a high-confidence way to reproduce the issue?

Not applicable as a feature PR rather than a broken existing-behavior report. Source inspection shows current main lacks the requested Cohere Command entries, and the PR head adds them.

Is this the best way to solve the issue?

Mostly yes: the Cohere plugin manifest, onboarding helper, provider entry, docs, and focused tests are the right owner boundary for this catalog refresh. The open question is the maintainable product/default decision, not a concrete patch defect.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against d27d7ba6d358.

Label changes

Label changes:

  • add proof: sufficient: Contributor real behavior proof is sufficient. Sufficient: the PR now includes copied live output for a GCP Crabbox Cohere Command A+ run with HTTP 200 and prompt/tool/image probes passing; private values should stay redacted in any follow-up logs.
  • add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🐚 platinum hermit and patch quality is 🐚 platinum hermit.
  • add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (live_output): Sufficient: the PR now includes copied live output for a GCP Crabbox Cohere Command A+ run with HTTP 200 and prompt/tool/image probes passing; private values should stay redacted in any follow-up logs.
  • remove rating: 🧂 unranked krab: Current PR rating is rating: 🐚 platinum hermit, so this older rating label is no longer current.
  • remove status: 📣 needs proof: Current PR status label is status: 👀 ready for maintainer look.

Label justifications:

  • P2: This is a provider-specific model catalog/default improvement with user-visible setup impact but limited blast radius.
  • merge-risk: 🚨 compatibility: Changing the Cohere onboarding default and curated live-sweep refs can affect fresh setup behavior and operator validation expectations.
  • merge-risk: 🚨 auth-provider: The PR changes model-choice metadata and default routing for the direct Cohere provider, which is part of provider/auth setup behavior.
  • rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🐚 platinum hermit and patch quality is 🐚 platinum hermit.
  • status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (live_output): Sufficient: the PR now includes copied live output for a GCP Crabbox Cohere Command A+ run with HTTP 200 and prompt/tool/image probes passing; private values should stay redacted in any follow-up logs.
  • proof: sufficient: Contributor real behavior proof is sufficient. Sufficient: the PR now includes copied live output for a GCP Crabbox Cohere Command A+ run with HTTP 200 and prompt/tool/image probes passing; private values should stay redacted in any follow-up logs.
Evidence reviewed

PR surface:

Source +107, Tests +81, Docs +7. Total +195 across 13 files.

View PR surface stats
Area Files Added Removed Net
Source 5 112 5 +107
Tests 4 88 7 +81
Docs 4 23 16 +7
Config 0 0 0 0
Generated 0 0 0 0
Other 0 0 0 0
Total 13 223 28 +195

What I checked:

Likely related people:

  • steipete: Recent merged provider catalog work touched the Cohere catalog, and this PR/linked issue are also authored and assigned to this login. (role: recent area contributor and likely follow-up owner; confidence: high; commits: e7492c6fca23, 80537c1ba408, eae3fc1a79dd; files: extensions/cohere/openclaw.plugin.json, docs/providers/cohere.md, src/agents/live-model-filter.ts)
  • vincentkoc: GitHub commit history shows this login authored the initial Cohere provider plugin and externalization work, including the central Cohere plugin files. (role: original provider plugin author; confidence: medium; commits: 85d5d9451977, c558e918c206, 5e7a0b155808; files: extensions/cohere/index.ts, extensions/cohere/openclaw.plugin.json, extensions/cohere/onboard.ts)
  • frank-beans: Recent provider-onboarding work preserved provider settings during onboarding updates, which is relevant to judging the default-model compatibility risk. (role: adjacent upgrade-safety contributor; confidence: medium; commits: 664464c750c0; files: src/plugin-sdk/provider-onboard.ts)
What the crustacean ranks mean
  • 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
  • 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
  • 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
  • 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
  • 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
  • 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
  • 🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

How this review workflow works
  • ClawSweeper keeps one durable marker-backed review comment per issue or PR.
  • Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
  • A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
  • PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
  • Maintainers can also comment @clawsweeper review to request a fresh review only.
  • Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
  • Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
  • Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.
Review history (1 earlier review cycle)
  • reviewed 2026-07-09T09:04:22.963Z sha fd244c2 :: needs real behavior proof before merge. :: [P3] Remove the release-owned changelog edit

@clawsweeper clawsweeper Bot added rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask. P2 Normal backlog priority with limited blast radius. merge-risk: 🚨 compatibility 🚨 May break existing users, config, migrations, defaults, or upgrade paths. merge-risk: 🚨 auth-provider 🚨 May break OAuth, tokens, provider routing, model choice, or credentials. labels Jul 9, 2026
@steipete
steipete marked this pull request as ready for review July 9, 2026 09:53
@steipete

steipete commented Jul 9, 2026

Copy link
Copy Markdown
Contributor Author

Land-ready proof for fd244c2:

Known proof gaps: none for the changed Cohere catalog/default/tool-capability path.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: fd244c29f7

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread CHANGELOG.md Outdated

### Changes

- **Cohere models:** add Command A+, Command A Reasoning, and Command A Vision to the direct provider, make Command A+ the onboarding default, and include it in modern live-model sweeps.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Remove the normal feature entry from CHANGELOG

For this non-release Cohere model change, adding an Unreleased changelog bullet bypasses the repo's release-generation contract (AGENTS.md says CHANGELOG.md is release-only and owned by release generation). Leaving this here can make the next generated changelog duplicate or preserve an entry that was not derived from the merged PR/direct-commit release sweep; put the release-note context in the PR/commit metadata instead.

Useful? React with 👍 / 👎.

@clawsweeper clawsweeper Bot added proof: sufficient ClawSweeper judged the real behavior proof convincing. rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR. and removed rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask. labels Jul 9, 2026
@steipete
steipete merged commit fc5a4de into main Jul 9, 2026
123 of 131 checks passed
@steipete
steipete deleted the codex/cohere-command-refresh branch July 9, 2026 10:07
@steipete

steipete commented Jul 9, 2026

Copy link
Copy Markdown
Contributor Author

Merged via squash.

github-actions Bot pushed a commit to Desicool/openclaw that referenced this pull request Jul 10, 2026
* feat(cohere): add current Command models

* docs: keep release notes out of PR
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

agents Agent runtime and tooling docs Improvements or additions to documentation gateway Gateway runtime maintainer Maintainer-authored PR merge-risk: 🚨 auth-provider 🚨 May break OAuth, tokens, provider routing, model choice, or credentials. merge-risk: 🚨 compatibility 🚨 May break existing users, config, migrations, defaults, or upgrade paths. P2 Normal backlog priority with limited blast radius. proof: sufficient ClawSweeper judged the real behavior proof convincing. rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. size: M status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Feature]: Add current Cohere Command models

1 participant