fix(anthropic): preserve configured Claude image capability by TurboTheTurtle · Pull Request #84180 · openclaw/openclaw

TurboTheTurtle · 2026-05-19T14:53:12Z

Summary

route configured/default models list rows through provider-owned resolved-model normalization before displaying model capabilities
keep Anthropic Claude 4 image capability policy in extensions/anthropic instead of duplicating Claude prefix data in core
cover the anthropic/claude-sonnet-4-5 configured/default path while preserving generic configured-row fallback behavior
add changelog entry

Real behavior proof

Behavior addressed: A configured default model such as agents.defaults.model = "anthropic/claude-sonnet-4-5" could still resolve through a stale text-only configured/catalog path, causing image input to be treated as unsupported even after the provider-discovered row fix.
Real environment tested: Local OpenClaw checkout on macOS, PR head 3bc1284add, using an isolated OPENCLAW_HOME=/private/tmp/openclaw-83756-proof and OPENCLAW_CONFIG_PATH=/private/tmp/openclaw-83756-proof/.openclaw/openclaw.json with only agents.defaults.model set to anthropic/claude-sonnet-4-5.

Exact steps or command run after the patch:

env OPENCLAW_HOME=/private/tmp/openclaw-83756-proof \
  OPENCLAW_CONFIG_PATH=/private/tmp/openclaw-83756-proof/.openclaw/openclaw.json \
  PATH=/Users/andy/.cache/codex-runtimes/codex-primary-runtime/dependencies/node/bin:/opt/homebrew/bin:/usr/bin:/bin:/usr/sbin:/sbin \
  pnpm openclaw models list --json

Evidence after fix: Terminal output from the real OpenClaw CLI after the patch:

{
  "count": 1,
  "models": [
    {
      "key": "anthropic/claude-sonnet-4-5",
      "name": "claude-sonnet-4-5",
      "input": "text+image",
      "contextWindow": 200000,
      "local": false,
      "available": false,
      "tags": [
        "default"
      ],
      "missing": false
    }
  ]
}

Observed result after fix: The configured default Anthropic Claude 4 row now reports input: "text+image" instead of text-only. The configured/default list row reaches the existing Anthropic provider-owned normalizeResolvedModel hook, so the same plugin-owned capability policy is used for this stale catalog path.
What was not tested: I did not make a live Anthropic API request. This proof covers the local configured model capability resolution and CLI-visible configured/default path that the reporter identified; the live provider call still depends on valid Anthropic credentials.

Validation

pnpm test src/agents/pi-embedded-runner/model.inline-provider.test.ts src/commands/models.list.e2e.test.ts
pnpm exec oxfmt --check --threads=1 src/agents/pi-embedded-runner/model.inline-provider.ts src/agents/pi-embedded-runner/model.inline-provider.test.ts src/commands/models/list.rows.ts src/commands/models/list.list-command.ts src/commands/models.list.e2e.test.ts CHANGELOG.md
git diff --check

Follow-up for #83756 (comment).

clawsweeper · 2026-05-19T14:54:25Z

Codex review: passed.

Workflow note: Future ClawSweeper reviews update this same comment in place.

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.

Summary
The PR routes model-list row construction through provider-owned resolved-model normalization for configured/default and catalog rows, adds Anthropic regression coverage, updates focused test mocks/fixtures, and adds a changelog entry.

Reproducibility: yes. at source level: current main renders configured/default list rows without calling the provider normalizer, while the Anthropic hook is where Claude 4 image capability is restored. The PR body also supplies terminal output showing the fixed configured row as text+image.

PR rating
Overall: 🐚 platinum hermit
Proof: 🦞 diamond lobster
Patch quality: 🐚 platinum hermit
Summary: Strong terminal proof and a focused implementation make this a good normal PR pending maintainer/automerge gates.

Rank-up moves:

none

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.

Real behavior proof
Sufficient (terminal): Sufficient terminal proof: the PR body shows a real local openclaw models list --json run after the patch returning the configured Anthropic Claude row with input: "text+image"; later commits are test and merge-expectation updates.

Next step before merge
No repair lane is needed because there are no blocking findings and the remaining action is maintainer/automerge handling of the exact PR head.

Security
Cleared: Cleared: the diff changes TypeScript model-list normalization, tests, workflow expectation tests, and changelog text without new dependencies, downloads, secrets, or privileged CI behavior.

Review details

Best possible solution:

Land the provider-owned normalization path so models list reports configured Claude capabilities from the Anthropic plugin, with exact-head CI and normal maintainer gates deciding merge.

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

Yes, at source level: current main renders configured/default list rows without calling the provider normalizer, while the Anthropic hook is where Claude 4 image capability is restored. The PR body also supplies terminal output showing the fixed configured row as text+image.

Is this the best way to solve the issue?

Yes. Reusing normalizeResolvedModel is the maintainable fix because it keeps Anthropic-specific capability policy in the provider plugin instead of duplicating Claude model-prefix rules in core list rendering.

Label changes:

add status: 🚀 automerge armed: This PR is in ClawSweeper's automerge lane. Sufficient (terminal): Sufficient terminal proof: the PR body shows a real local openclaw models list --json run after the patch returning the configured Anthropic Claude row with input: "text+image"; later commits are test and merge-expectation updates.
remove status: 👀 ready for maintainer look: Current PR status label is status: 🚀 automerge armed.

What I checked:

Current main gap: appendConfiguredRows on current main resolves or falls back to a configured model and renders it directly with toModelRow, so configured/default rows can miss provider-owned normalizeResolvedModel capability fixes. (src/commands/models/list.rows.ts:490, 9108ae011471)
Existing provider-owned policy: The Anthropic plugin owns Claude 4 image capability normalization: applyAnthropicImageInputCapability adds input: ["text", "image"] for modern Claude model refs and normalizeResolvedModel composes that with the Opus context-window fix. (extensions/anthropic/register.runtime.ts:389, 9108ae011471)
Runtime resolver precedent: The embedded runner already invokes normalizeProviderResolvedModelWithPlugin while resolving models, so the PR aligns list-row display with the existing runtime model-resolution seam instead of adding Anthropic prefixes to core. (src/agents/pi-embedded-runner/model.ts:256, 9108ae011471)
PR diff reviewed: The PR adds normalizeListRowWithProviderPlugin, threads workspaceDir into row context, applies provider normalization before list-row rendering/suppression, adds configured Claude e2e coverage, and keeps inline-provider input parsing from duplicating Anthropic policy. (src/commands/models/list.rows.ts:119, 7a1caa7dff8e)
Related merged provenance: The linked predecessor fix(anthropic): preserve Claude image capability #83756 merged as 98cc6df7ff2332f874d55b1fbb66dbeb09609b52 on 2026-05-18 and introduced the Anthropic provider-owned image-capability normalization that this PR extends to configured/default model-list rows. (extensions/anthropic/register.runtime.ts:405, 98cc6df7ff23)
Proof and exact-head checks: The PR body supplies terminal output from a real openclaw models list --json run showing anthropic/claude-sonnet-4-5 as text+image; public check-run summary for head 7a1caa7dff8e showed 81 success, 36 skipped, 1 neutral, and one cancelled auto-response check. (7a1caa7dff8e)

Likely related people:

TurboTheTurtle: Authored the merged related Anthropic image-capability PR and this follow-up explicitly extends that provider-owned normalization path to the model-list configured/default surface. (role: recent feature-history contributor; confidence: high; commits: 98cc6df7ff23, 7a1caa7dff8e; files: extensions/anthropic/register.runtime.ts, src/commands/models/list.rows.ts, src/commands/models.list.e2e.test.ts)

Codex review notes: model gpt-5.5, reasoning high; reviewed against ad925bd43b1e.

TurboTheTurtle · 2026-05-19T15:24:00Z

@clawsweeper re-review

clawsweeper · 2026-05-19T15:25:44Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Complete
Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26107163778
Updated: 2026-05-19T15:34:19.301Z

TurboTheTurtle · 2026-05-19T21:25:15Z

Pushed 6770905bec to fix the current checks-node-agentic-commands-models failure.

Root cause: the forward-compat model-list test mocks ../../agents/agent-scope.js, but newer model runtime policy code now imports resolveSessionAgentIds and listAgentEntries from that module. The production code was fine; the test mock was stale.

Validation on PR head 6770905bec:

pnpm exec vitest run --config test/vitest/vitest.commands.config.ts src/commands/models/list.list-command.forward-compat.test.ts — 27 passed
workflow-matched agentic-commands-models include shard via OPENCLAW_VITEST_INCLUDE_FILE + scripts/test-projects.mjs — 18 files / 199 tests passed
pnpm exec oxfmt --check --threads=1 src/commands/models/list.list-command.forward-compat.test.ts
git diff --check

@clawsweeper re-review

clawsweeper · 2026-05-19T23:47:52Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Complete
Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26132281488
Updated: 2026-05-19T23:54:23.967Z

…gured-claude-image-capability

clawsweeper · 2026-05-19T23:54:17Z

ClawSweeper PR egg

✨ Hatched: 🥚 common Pearl Review Wisp

Hatch command

Comment @clawsweeper hatch when this PR is hatchable.

Hatchability rules:

Merged PRs are hatchable.
Open PRs are hatchable when they are status: 👀 ready for maintainer look, status: 🚀 automerge armed, or labeled clawsweeper:automerge.
Closed unmerged PRs are hatchable only when one of those hatchable labels is still present in the durable record.

Rarity: 🥚 common.
Trait: sleeps inside passing CI.
Image traits: location CI tidepool; accessory lint brush; palette seafoam, black, and opal; mood curious; pose sitting proudly on a smooth stone; shell matte ceramic shell; lighting calm overcast light; background small green status lights.
Share on X: post this hatch
Copy: My PR egg hatched a 🥚 common Pearl Review Wisp in ClawSweeper.

What is this egg doing here?

Eggs appear after the PR passes real-behavior proof. It is here for vibes, not verdicts: it does not change labels, ratings, merge decisions, or automation.
The shell reacts to review momentum: open follow-up work warms it up, re-review makes it wobble, and a clean final review lets it hatch.
Hatchability usually comes from sufficient real-behavior proof, no blocking P0/P1/P2 findings, no security attention needed, and clean correctness. A merged PR is already final, so merge makes the egg hatchable independently.
The hatch is seeded from this repository and PR number, so the same PR keeps the same creature; the reviewed head SHA can only change safe visual details.
Rarity is just collectible sparkle: 🥚 common, 🌱 uncommon, 💎 rare, ✨ glimmer, and 🌈 legendary.

TurboTheTurtle · 2026-05-19T23:55:22Z

Pushed a follow-up for the new build-artifacts failure. The branch was behind current upstream/main, and the failing core-support-boundary tests were workflow/tooling expectations for updated Tideclaw alpha concurrency rules. I merged current upstream/main into the PR branch and updated the two test expectations to match the current workflow YAML.

Local validation:

node scripts/run-vitest.mjs test/scripts/package-acceptance-workflow.test.ts test/scripts/plugin-prerelease-test-plan.test.ts (33 tests passed)
./node_modules/.bin/oxfmt --check --threads=1 test/scripts/package-acceptance-workflow.test.ts test/scripts/plugin-prerelease-test-plan.test.ts
git diff --check

Previous validation still stands for the model fixture fix:

node scripts/run-tsgo.mjs -p test/tsconfig/tsconfig.core.test.json --incremental --tsBuildInfoFile .artifacts/tsgo-cache/core-test.tsbuildinfo
focused model Vitest suite (42 tests passed)

@clawsweeper re-review

clawsweeper · 2026-05-19T23:57:09Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Complete
Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26132614924
Updated: 2026-05-20T00:05:08.826Z

Takhoffman · 2026-05-20T03:22:13Z

@clawsweeper automerge

clawsweeper · 2026-05-20T03:24:08Z

🦞✅
ClawSweeper merged this PR after the passing review.

Source: clawsweeper[bot]
Feedback: structured ClawSweeper verdict: pass (sha=7a1caa7dff8ee0e5ee69e4e3a22b867d27115f9e)
Merge status: merged by ClawSweeper automerge
Merged at: 2026-05-20T03:32:14Z
Merge commit: 33fc2375f8a3

What merged:

The PR routes model-list row construction through provider-owned resolved-model normalization for configured ... rows, adds Anthropic regression coverage, updates focused test mocks/fixtures, and adds a changelog entry.
Reproducibility: yes. at source level: current main renders configured/default list rows without calling the ... ty is restored. The PR body also supplies terminal output showing the fixed configured row as text+image.

Automerge notes:

PR branch already contained follow-up commit before automerge: test(models): update forward compat agent-scope mock
PR branch already contained follow-up commit before automerge: test(models): isolate provider catalog row tests
PR branch already contained follow-up commit before automerge: test(models): complete provider catalog fixtures
PR branch already contained follow-up commit before automerge: Merge remote-tracking branch 'upstream/main' into fix/anthropic-confi…
PR branch already contained follow-up commit before automerge: test(workflows): match alpha concurrency rules

The automerge loop is complete.

Automerge progress:

2026-05-20 03:24:07 UTC review queued 7a1caa7dff8e (queued)

2026-05-20 03:31:56 UTC review passed 7a1caa7dff8e (structured ClawSweeper verdict: pass (sha=7a1caa7dff8ee0e5ee69e4e3a22b867d27115...)

2026-05-20 03:32:16 UTC merged 7a1caa7dff8e (merged by ClawSweeper automerge)

Summary: - The PR routes model-list row construction through provider-owned resolved-model normalization for configured ... rows, adds Anthropic regression coverage, updates focused test mocks/fixtures, and adds a changelog entry. - Reproducibility: yes. at source level: current main renders configured/default list rows without calling the ... ty is restored. The PR body also supplies terminal output showing the fixed configured row as `text+image`. Automerge notes: - PR branch already contained follow-up commit before automerge: test(models): update forward compat agent-scope mock - PR branch already contained follow-up commit before automerge: test(models): isolate provider catalog row tests - PR branch already contained follow-up commit before automerge: test(models): complete provider catalog fixtures - PR branch already contained follow-up commit before automerge: Merge remote-tracking branch 'upstream/main' into fix/anthropic-confi… - PR branch already contained follow-up commit before automerge: test(workflows): match alpha concurrency rules Validation: - ClawSweeper review passed for head 7a1caa7dff8ee0e5ee69e4e3a22b867d27115f9e. - Required merge gates passed before the squash merge. Prepared head SHA: 7a1caa7dff8ee0e5ee69e4e3a22b867d27115f9e Review: openclaw/openclaw#84180 (comment) Co-authored-by: Andy Ye <35905412+TurboTheTurtle@users.noreply.github.com> Co-authored-by: clawsweeper[bot] <274271284+clawsweeper[bot]@users.noreply.github.com> Approved-by: takhoffman Co-authored-by: takhoffman <781889+takhoffman@users.noreply.github.com>

…#84180) Summary: - The PR routes model-list row construction through provider-owned resolved-model normalization for configured ... rows, adds Anthropic regression coverage, updates focused test mocks/fixtures, and adds a changelog entry. - Reproducibility: yes. at source level: current main renders configured/default list rows without calling the ... ty is restored. The PR body also supplies terminal output showing the fixed configured row as `text+image`. Automerge notes: - PR branch already contained follow-up commit before automerge: test(models): update forward compat agent-scope mock - PR branch already contained follow-up commit before automerge: test(models): isolate provider catalog row tests - PR branch already contained follow-up commit before automerge: test(models): complete provider catalog fixtures - PR branch already contained follow-up commit before automerge: Merge remote-tracking branch 'upstream/main' into fix/anthropic-confi… - PR branch already contained follow-up commit before automerge: test(workflows): match alpha concurrency rules Validation: - ClawSweeper review passed for head 7a1caa7. - Required merge gates passed before the squash merge. Prepared head SHA: 7a1caa7 Review: openclaw#84180 (comment) Co-authored-by: Andy Ye <35905412+TurboTheTurtle@users.noreply.github.com> Co-authored-by: clawsweeper[bot] <274271284+clawsweeper[bot]@users.noreply.github.com> Approved-by: takhoffman Co-authored-by: takhoffman <781889+takhoffman@users.noreply.github.com>

TurboTheTurtle mentioned this pull request May 19, 2026

fix(anthropic): preserve Claude image capability #83756

Merged

openclaw-barnacle Bot added agents Agent runtime and tooling size: S triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. labels May 19, 2026

clawsweeper Bot added the rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. label May 19, 2026

TurboTheTurtle force-pushed the fix/anthropic-configured-claude-image-capability branch from 3fac142 to be8789b Compare May 19, 2026 14:59

clawsweeper Bot added status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask. P2 Normal backlog priority with limited blast radius. labels May 19, 2026

openclaw-barnacle Bot added commands Command implementations proof: supplied External PR includes structured after-fix real behavior proof. and removed triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. labels May 19, 2026

fix(anthropic): preserve configured Claude image capability

3bc1284

TurboTheTurtle force-pushed the fix/anthropic-configured-claude-image-capability branch from be8789b to 3bc1284 Compare May 19, 2026 15:21

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 19, 2026

test(models): update forward compat agent-scope mock

6770905

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 19, 2026

Merge remote-tracking branch 'upstream/main' into fix/anthropic-confi…

ddf9aa1

…gured-claude-image-capability

clawsweeper Bot added the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 19, 2026

test(workflows): match alpha concurrency rules

7a1caa7

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 19, 2026

clawsweeper Bot added the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 20, 2026

clawsweeper Bot added the clawsweeper:automerge Maintainer opted this PR into bounded ClawSweeper-reviewed automerge label May 20, 2026

clawsweeper Bot added status: 🚀 automerge armed This PR is in ClawSweeper's automerge lane. and removed status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR. labels May 20, 2026

clawsweeper Bot merged commit 33fc237 into openclaw:main May 20, 2026
127 of 130 checks passed

TurboTheTurtle deleted the fix/anthropic-configured-claude-image-capability branch May 20, 2026 03:39

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix(anthropic): preserve configured Claude image capability#84180

fix(anthropic): preserve configured Claude image capability#84180
clawsweeper[bot] merged 6 commits into
openclaw:mainfrom
TurboTheTurtle:fix/anthropic-configured-claude-image-capability

TurboTheTurtle commented May 19, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented May 19, 2026 •

edited

Loading

Uh oh!

TurboTheTurtle commented May 19, 2026

Uh oh!

clawsweeper Bot commented May 19, 2026 •

edited

Loading

Uh oh!

TurboTheTurtle commented May 19, 2026

Uh oh!

clawsweeper Bot commented May 19, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented May 19, 2026 •

edited

Loading

Uh oh!

TurboTheTurtle commented May 19, 2026

Uh oh!

clawsweeper Bot commented May 19, 2026 •

edited

Loading

Uh oh!

Takhoffman commented May 20, 2026

Uh oh!

clawsweeper Bot commented May 20, 2026 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

TurboTheTurtle commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Real behavior proof

Validation

Uh oh!

clawsweeper Bot commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

TurboTheTurtle commented May 19, 2026

Uh oh!

clawsweeper Bot commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

TurboTheTurtle commented May 19, 2026

Uh oh!

clawsweeper Bot commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

clawsweeper Bot commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Hatch command

Uh oh!

TurboTheTurtle commented May 19, 2026

Uh oh!

clawsweeper Bot commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Takhoffman commented May 20, 2026

Uh oh!

clawsweeper Bot commented May 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

TurboTheTurtle commented May 19, 2026 •

edited

Loading

clawsweeper Bot commented May 19, 2026 •

edited

Loading

clawsweeper Bot commented May 19, 2026 •

edited

Loading

clawsweeper Bot commented May 19, 2026 •

edited

Loading

clawsweeper Bot commented May 19, 2026 •

edited

Loading

clawsweeper Bot commented May 19, 2026 •

edited

Loading

clawsweeper Bot commented May 20, 2026 •

edited

Loading