fix(agents): avoid eager alias normalization for default models by TurboTheTurtle · Pull Request #86671 · openclaw/openclaw

TurboTheTurtle · 2026-05-25T23:36:43Z

Summary

avoid building a full model alias index when resolving a configured default model
skip model key parsing for configured model entries that do not define an alias
keep slash-form configured aliases working while making provider-qualified defaults resolve through the direct parse path

Real behavior proof

Behavior addressed: provider-qualified default model resolution no longer walks and normalizes every configured model just to resolve openai/gpt-5.5 on the inbound reply hot path.
Real environment tested: local OpenClaw source checkout on macOS at current PR head c8506316c7c05748c3389b5a4c4c47344579519b, using Node v24.15.0 with tsx to execute the production model-selection source directly.
Exact steps or command run after this patch: node --import tsx - <<'EOF' executed resolveConfiguredModelRef and buildModelAliasIndex from ./src/agents/model-selection.ts against a synthetic config with 97 aliasless configured models plus two configured aliases, then asserted that the provider-qualified default resolved to openai/gpt-5.5, the alias index contained only the two real aliases, and the slash-form alias preserved the routed model path. The proof data uses synthetic model names only.
Evidence after fix: Copied console output from the real command:

{
  "head": "c8506316c7c05748c3389b5a4c4c47344579519b",
  "scenario": "provider-qualified default model with many aliasless configured models",
  "configuredModelCount": 99,
  "resolved": {
    "provider": "openai",
    "model": "gpt-5.5"
  },
  "aliasCount": 2,
  "miniAlias": {
    "provider": "openai",
    "model": "gpt-5.5-mini"
  },
  "slashAlias": {
    "provider": "openrouter",
    "model": "xiaomi/mimo-v2-pro-mit"
  },
  "elapsedMs": 92.697
}
PASS current-head provider-qualified default resolved without indexing aliasless configured models

Observed result after fix: resolveConfiguredModelRef resolved the provider-qualified default directly to openai/gpt-5.5; buildModelAliasIndex ignored the 97 aliasless configured model entries and returned only the two configured aliases; the routed slash-form alias still resolved to openrouter/xiaomi/mimo-v2-pro-mit.
What was not tested: a live WhatsApp/gateway workload with the reporter's exact provider plugin normalization latency; the local proof exercises the production source hot path and regression shape from #86635 without private user data.

Verification

node --import tsx - <<'EOF' current-head production model-selection proof above
node scripts/run-vitest.mjs src/agents/model-selection.test.ts
node scripts/run-vitest.mjs src/agents/tools/message-tool.test.ts --testNamePattern 'uses a non-webchat session key'
pnpm exec oxfmt --check --threads=1 src/agents/model-selection.test.ts
git log --format='%h %an <%ae> %s' upstream/main..HEAD

Fixes #86635

Co-authored-by: Andy Ye 35905412+TurboTheTurtle@users.noreply.github.com

clawsweeper · 2026-05-25T23:43:55Z

Codex review: needs maintainer review before merge. Reviewed May 25, 2026, 11:02 PM ET / 03:02 UTC.

Summary
This PR adds model-selection regression tests that mock provider model normalization and assert aliasless configured models are not normalized while provider-qualified defaults still resolve directly.

PR surface: Tests +61. Total +61 across 1 file.

Reproducibility: yes. the linked bug gives a concrete source path and timing evidence, and current source inspection shows the direct-resolution fix path. I did not run a failing current-main repro because main already contains the lazy/direct-resolution implementation.

Review metrics: none identified.

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:

none

Next step before merge
No repair lane is needed; the patch has no actionable code defect and the remaining action is normal PR review and CI handling.

Security
Cleared: The diff is test-only and does not touch dependencies, workflows, credentials, package resolution, or runtime security boundaries.

Review details

Best possible solution:

Land the focused regression tests after normal CI and maintainer review, then close the linked bug when the implementation plus coverage are on main.

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

Yes: the linked bug gives a concrete source path and timing evidence, and current source inspection shows the direct-resolution fix path. I did not run a failing current-main repro because main already contains the lazy/direct-resolution implementation.

Is this the best way to solve the issue?

Yes: as a test-only follow-up, the PR checks the hot-path invariants in the existing model-selection unit file without adding new runtime behavior or config surface.

AGENTS.md: found and applied where relevant.

Codex review notes: model gpt-5.5, reasoning high; reviewed against c9d0464ed116.

Label changes

Label changes:

add proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes current-head terminal proof from executing the production model-selection source against a synthetic catalog and observing the expected direct resolution and alias filtering.
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 (terminal): The PR body includes current-head terminal proof from executing the production model-selection source against a synthetic catalog and observing the expected direct resolution and alias filtering.
remove rating: 🦪 silver shellfish: 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:

P1: The PR is tied to a high-impact agent/gateway event-loop starvation bug that can delay inbound replies and channel health checks for real users.
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 (terminal): The PR body includes current-head terminal proof from executing the production model-selection source against a synthetic catalog and observing the expected direct resolution and alias filtering.
proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes current-head terminal proof from executing the production model-selection source against a synthetic catalog and observing the expected direct resolution and alias filtering.

Evidence reviewed

PR surface:

Tests +61. Total +61 across 1 file.

View PR surface stats

Area	Files	Added	Removed	Net
Source	0	0	0	0
Tests	1	62	1	+61
Docs	0	0	0	0
Config	0	0	0	0
Generated	0	0	0	0
Other	0	0	0	0
Total	1	62	1	+61

What I checked:

PR diff adds call-count coverage for provider normalization: The PR head adds a hoisted provider-normalization mock and asserts buildModelAliasIndex only calls runtime normalization for the configured alias entry, not the 25 aliasless entries. (src/agents/model-selection.test.ts:100, c8506316c7c0)
PR diff covers provider-qualified default resolution: The PR head adds a regression test where primary model openai/gpt-5.5 resolves while 25 aliasless configured models do not trigger per-entry provider runtime normalization. (src/agents/model-selection.test.ts:1808, c8506316c7c0)
Current source skips aliasless configured-model rows: Current main builds alias candidates from entries with a non-empty alias, so aliasless defaults are filtered before model-key parsing and runtime normalization. (src/agents/model-selection-shared.ts:107, c9d0464ed116)
Current source uses direct provider-qualified parsing after alias scan: resolveConfiguredModelRef now scans configured aliases directly and then resolves slash-form provider/model strings without first constructing the full alias index. (src/agents/model-selection-shared.ts:663, c9d0464ed116)
Runtime call path matches the linked hot path: Inbound replies still call resolveDefaultModel before agent dispatch, and that helper resolves the default model then builds the alias index for later directive handling, matching the linked bug surface. (src/auto-reply/reply/get-reply.ts:252, c9d0464ed116)
Implementation provenance: Commit ea2496b introduced the lazy manifest context, alias-candidate filtering, and direct default resolution behavior that this PR tests. (src/agents/model-selection-shared.ts:488, ea2496b00c5c)

Likely related people:

Deepflame: git show and git blame tie the current lazy manifest/alias-resolution implementation in model-selection-shared.ts to ea2496b. (role: recent area contributor; confidence: high; commits: ea2496b00c5c; files: src/agents/model-selection-shared.ts, src/agents/model-selection.test.ts)
@vincentkoc: git blame and recent history show 84929e4 carried the surrounding model-selection shared/test baseline immediately before the lazy-resolution change. (role: recent adjacent contributor; confidence: medium; commits: 84929e426506; files: src/agents/model-selection-shared.ts, src/agents/model-selection.ts, src/agents/model-selection.test.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.

clawsweeper · 2026-05-25T23:59:46Z

ClawSweeper PR egg

✨ Hatched: 🥚 common Neon Patch Peep

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: polishes edge cases.
Image traits: location flaky test forest; accessory miniature diff map; palette pearl, teal, and neon green; mood mischievous; pose sitting proudly on a smooth stone; shell matte ceramic shell; lighting subtle sparkle highlights; background miniature CI buoys.
Share on X: post this hatch
Copy: My PR egg hatched a 🥚 common Neon Patch Peep 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 · 2026-05-26T02:56:42Z

@clawsweeper re-review

clawsweeper · 2026-05-26T02:56:45Z

🦞🧹
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/26429647938
Updated: 2026-05-26T03:07:01.833Z

steipete · 2026-05-27T04:46:59Z

Pre-merge verification for #86671:

Behavior addressed: regression coverage for provider-qualified default model resolution avoiding eager alias work across aliasless configured models.
Real environment tested: local OpenClaw checkout on main after syncing origin/main, plus PR head CI.
Exact steps or command run after this patch:

node scripts/run-vitest.mjs src/agents/model-selection.test.ts
node --import tsx -e "const {resolveConfiguredModelRef, buildModelAliasIndex}=await import('./src/agents/model-selection.ts'); const models=Object.fromEntries(Array.from({length:97},(_,i)=>['anthropic/claude-extra-'+i,{}])); const cfg={agents:{defaults:{model:{primary:'openai/gpt-5.5'},models:{...models,'openai/gpt-5.5-mini':{alias:'mini'},'openrouter/xiaomi/mimo-v2-pro-mit':{alias:'router/xiaomi'}}}}}; const resolved=resolveConfiguredModelRef({cfg,defaultProvider:'anthropic',defaultModel:'claude-sonnet-4-6'}); const idx=buildModelAliasIndex({cfg,defaultProvider:'anthropic'}); console.log(JSON.stringify({resolved,aliasCount:idx.byAlias.size,mini:idx.byAlias.get('mini')?.ref,slashAlias:idx.byAlias.get('router/xiaomi')?.ref},null,2)); if(resolved.provider!=='openai'||resolved.model!=='gpt-5.5'||idx.byAlias.size!==2) process.exit(1);"
git diff --check origin/main...refs/remotes/pull/86671 -- src/agents/model-selection.test.ts
Evidence after fix: Vitest passed 206 tests across 2 files; inline production-path proof resolved openai/gpt-5.5 and kept only 2 aliases with 97 aliasless configured models; diff whitespace check was clean.
Observed result after fix: the current production path already resolves the provider-qualified default directly and the PR adds focused regression coverage for that behavior.
CI: gh pr checks 86671 shows relevant required lanes passing, including CI, check-lint, check-prod-types, check-test-types, checks-node-agentic-agents, security-fast, OpenGrep, CodeQL/security selected lanes, and Real behavior proof. Routine auto-response/label/docs/non-selected lanes were ignored or skipped.
What was not tested: no live gateway/WhatsApp workload; this PR is test-only and current production behavior was proven locally on the model-selection path.

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 25, 2026

openclaw-barnacle Bot added 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 25, 2026

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 26, 2026

TurboTheTurtle force-pushed the fix/provider-default-alias-fast-path-86635 branch 2 times, most recently from 7861208 to e3802f4 Compare May 26, 2026 00:21

fix(agents): avoid eager alias normalization for default models

c850631

TurboTheTurtle force-pushed the fix/provider-default-alias-fast-path-86635 branch from e3802f4 to c850631 Compare May 26, 2026 00:43

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 26, 2026

clawsweeper Bot added proof: sufficient ClawSweeper judged the real behavior proof convincing. and removed rating: 🦪 silver shellfish Thin PR readiness signal; proof, validation, or implementation needs work. labels May 26, 2026

steipete self-assigned this May 27, 2026

steipete merged commit 2e3b4b5 into openclaw:main May 27, 2026
150 of 161 checks passed

steipete mentioned this pull request May 27, 2026

[Bug]: Provider-qualified default model resolution eagerly builds alias index and can block gateway event loop ~80s #86635

Closed

TurboTheTurtle deleted the fix/provider-default-alias-fast-path-86635 branch May 27, 2026 08:00

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix(agents): avoid eager alias normalization for default models#86671

fix(agents): avoid eager alias normalization for default models#86671
steipete merged 1 commit into
openclaw:mainfrom
TurboTheTurtle:fix/provider-default-alias-fast-path-86635

TurboTheTurtle commented May 25, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented May 25, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented May 25, 2026 •

edited

Loading

Uh oh!

TurboTheTurtle commented May 26, 2026

Uh oh!

clawsweeper Bot commented May 26, 2026 •

edited

Loading

Uh oh!

steipete commented May 27, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

TurboTheTurtle commented May 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Real behavior proof

Verification

Uh oh!

clawsweeper Bot commented May 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

clawsweeper Bot commented May 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Hatch command

Uh oh!

TurboTheTurtle commented May 26, 2026

Uh oh!

clawsweeper Bot commented May 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

steipete commented May 27, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

TurboTheTurtle commented May 25, 2026 •

edited

Loading

clawsweeper Bot commented May 25, 2026 •

edited

Loading

clawsweeper Bot commented May 25, 2026 •

edited

Loading

clawsweeper Bot commented May 26, 2026 •

edited

Loading