Agents: prefer runtime-resolved metadata for explicit codex gpt-5.4 by ruclaw7 · Pull Request #62694 · openclaw/openclaw

ruclaw7 · 2026-04-07T20:30:43Z

Summary

Fixes openai-codex/gpt-5.4 resolution so an explicit model selection can still prefer provider runtime metadata when the provider reports a larger context window than the static registry entry.

This PR now:

lets the openai-codex plugin opt into runtime-preferred resolution via a provider-owned preferRuntimeResolvedModel hook instead of a hardcoded core special case
compares the explicit registry model with the provider runtime-resolved model only for providers that opt in, so normal explicit-model resolution still short-circuits for everyone else
preserves the correct runtime hook inputs by passing the configured workspace directory as workspaceDir while keeping agentDir available separately in context
adds focused regression coverage for the gpt-5.4 runtime-preference path
includes the follow-up review fixes that repaired the runtime hook plumbing and workspace handling

Testing

pnpm exec vitest run src/agents/pi-embedded-runner/model.test.ts
pnpm build

Fixes #55461

Thanks to Rudi Cilibrasi and Metagood.com for building and funding this fix.

greptile-apps · 2026-04-07T20:35:11Z

Greptile Summary

This PR makes resolveModelWithRegistry and resolveModelAsync prefer the plugin-dynamic openai-codex/gpt-5.4 model over the registry-resolved one when it reports a larger context window, and adds focused sync and async regression tests for the short-circuit path. All findings are P2.

Confidence Score: 5/5

Safe to merge — both findings are non-blocking style suggestions with no present runtime defect.

Both findings are P2: one is an architectural style concern about hardcoded provider/model names in core, and the other is a minor performance note about an unconditional hook call whose result is discarded for most callers. Neither affects correctness.

src/agents/pi-embedded-runner/model.ts — the unconditional plugin-dynamic-model resolution (line 570) and the hardcoded provider/model check (lines 529–534).

Prompt To Fix All With AI

This is a comment left during a code review.
Path: src/agents/pi-embedded-runner/model.ts
Line: 529-534

Comment:
**Hardcoded provider/model special case in core**

This function names `openai-codex` and `gpt-5.4` directly in core model resolution. The project's architecture rules prohibit hardcoded provider/model special cases in core when a manifest, capability, or plugin-owned contract can express the same behavior. A manifest-level flag such as `preferRuntimeContextWindow: true` or a plugin-owned hook would let the openai-codex plugin own this preference rather than requiring a named core carve-out for each model variant.

**Context Used:** CLAUDE.md ([source](https://app.greptile.com/review/custom-context?memory=fd949e91-5c3a-4ab5-90a1-cbe184fd6ce8))

How can I resolve this? If you propose a fix, please make it concise.

---

This is a comment left during a code review.
Path: src/agents/pi-embedded-runner/model.ts
Line: 570-573

Comment:
**Plugin dynamic hook fires for every explicitly-resolved model**

`resolvePluginDynamicModelWithRegistry` is called unconditionally before the `if (explicitModel?.kind === "resolved")` guard at line 571. For every provider whose model has an explicit registry entry but is not `openai-codex/gpt-5.4`, the hook fires and its result is immediately discarded. Guarding the call avoids the wasted `runProviderDynamicModel` invocation for all other providers.

```suggestion
  const pluginDynamicModel = shouldCompareOpenAICodexRuntimeResolvedModel(normalizedParams)
    ? resolvePluginDynamicModelWithRegistry(normalizedParams)
    : explicitModel?.kind === "resolved"
      ? undefined
      : resolvePluginDynamicModelWithRegistry(normalizedParams);
```

How can I resolve this? If you propose a fix, please make it concise.

_{Reviews (1): Last reviewed commit: "Agents: prefer runtime codex gpt-5.4 met..." | Re-trigger Greptile}

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 2084593780

ℹ️ About Codex in GitHub

Codex has been enabled to automatically 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 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 794570ffb9

ℹ️ About Codex in GitHub

Codex has been enabled to automatically 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 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 49ef268816

ℹ️ About Codex in GitHub

Codex has been enabled to automatically 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 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

ruclaw7 · 2026-04-08T01:04:57Z

Follow-up pass complete: I addressed the later Codex review note in 467fc20a28, resolved the remaining review thread, and pushed the update to the PR branch.

…ruclaw7

…ruclaw7)

obviyus

Reviewed latest changes; landing now.

obviyus · 2026-04-08T02:46:42Z

Landed on main.

Landed source commit: 78dfce3
Merge commit: 81969c7

Thanks @ruclaw7.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 78dfce34d7

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".

chatgpt-codex-connector · 2026-04-08T02:50:55Z

+    ) {
+      return explicitModel.model;
+    }
+    const pluginDynamicModel = resolvePluginDynamicModelWithRegistry(normalizedParams);


Pass workspaceDir into runtime model lookup

The new explicit-model runtime-comparison path calls resolvePluginDynamicModelWithRegistry(...) without propagating the workspace that was just used by shouldCompareProviderRuntimeResolvedModel(...). In this flow, shouldPreferProviderRuntimeResolvedModel is evaluated against cfg.agents.defaults.workspace, but runProviderDynamicModel still resolves provider plugins from the active global registry workspace when workspaceDir is omitted, so a workspace-installed provider can opt in to comparison and then have dynamic resolution run in a different workspace (or not run at all). Fresh evidence here is that this mismatch is now reachable from the new explicit branch at resolveModelWithRegistry, so explicit model comparisons can silently fall back to stale metadata in multi-workspace setups.

Useful? React with 👍 / 👎.

@ruclaw7

* Agents: prefer runtime codex gpt-5.4 metadata * Agents: move codex gpt-5.4 override into provider hook * fix: repair codex runtime preference hooks * fix: use workspace dir for codex runtime preference * test: cover codex workspace dir hook * fix: prefer codex gpt-5.4 runtime metadata (#62694) (thanks @ruclaw7) --------- Co-authored-by: Rudi Cilibrasi <cilibrar@gmail.com> Co-authored-by: Rudi Cilibrasi <rudi@metagood.com> Co-authored-by: Ayaan Zaidi <hi@obviy.us>

…ruclaw7

…ruclaw7) * Agents: prefer runtime codex gpt-5.4 metadata * Agents: move codex gpt-5.4 override into provider hook * fix: repair codex runtime preference hooks * fix: use workspace dir for codex runtime preference * test: cover codex workspace dir hook * fix: prefer codex gpt-5.4 runtime metadata (openclaw#62694) (thanks @ruclaw7) --------- Co-authored-by: Rudi Cilibrasi <cilibrar@gmail.com> Co-authored-by: Rudi Cilibrasi <rudi@metagood.com> Co-authored-by: Ayaan Zaidi <hi@obviy.us>

…ruclaw7

…ruclaw7) * Agents: prefer runtime codex gpt-5.4 metadata * Agents: move codex gpt-5.4 override into provider hook * fix: repair codex runtime preference hooks * fix: use workspace dir for codex runtime preference * test: cover codex workspace dir hook * fix: prefer codex gpt-5.4 runtime metadata (openclaw#62694) (thanks @ruclaw7) --------- Co-authored-by: Rudi Cilibrasi <cilibrar@gmail.com> Co-authored-by: Rudi Cilibrasi <rudi@metagood.com> Co-authored-by: Ayaan Zaidi <hi@obviy.us>

…ruclaw7

…ruclaw7) * Agents: prefer runtime codex gpt-5.4 metadata * Agents: move codex gpt-5.4 override into provider hook * fix: repair codex runtime preference hooks * fix: use workspace dir for codex runtime preference * test: cover codex workspace dir hook * fix: prefer codex gpt-5.4 runtime metadata (openclaw#62694) (thanks @ruclaw7) --------- Co-authored-by: Rudi Cilibrasi <cilibrar@gmail.com> Co-authored-by: Rudi Cilibrasi <rudi@metagood.com> Co-authored-by: Ayaan Zaidi <hi@obviy.us>

openclaw-barnacle Bot added agents Agent runtime and tooling size: S labels Apr 7, 2026

greptile-apps Bot reviewed Apr 7, 2026

View reviewed changes

Comment thread src/agents/pi-embedded-runner/model.ts Outdated

Comment thread src/agents/pi-embedded-runner/model.ts Outdated

chatgpt-codex-connector Bot reviewed Apr 7, 2026

View reviewed changes

Comment thread src/agents/pi-embedded-runner/model.ts Outdated

openclaw-barnacle Bot added the extensions: openai label Apr 7, 2026

ruclaw7 force-pushed the fix/codex-gpt54-resolution branch from 2c53002 to 794570f Compare April 7, 2026 23:53

chatgpt-codex-connector Bot reviewed Apr 7, 2026

View reviewed changes

Comment thread src/agents/pi-embedded-runner/model.ts

chatgpt-codex-connector Bot reviewed Apr 8, 2026

View reviewed changes

Comment thread src/agents/pi-embedded-runner/model.ts Outdated

ruclaw7 changed the title ~~Agents: prefer runtime metadata for codex gpt-5.4~~ Agents: prefer runtime-resolved metadata for explicit codex gpt-5.4 Apr 8, 2026

openclaw-barnacle Bot added size: M and removed size: S labels Apr 8, 2026

obviyus self-assigned this Apr 8, 2026

rudi-cilibrasi and others added 6 commits April 8, 2026 08:10

Agents: prefer runtime codex gpt-5.4 metadata

13ff07a

Agents: move codex gpt-5.4 override into provider hook

f881cec

fix: repair codex runtime preference hooks

cec6b75

fix: use workspace dir for codex runtime preference

c57c6e2

test: cover codex workspace dir hook

154d5fd

fix: prefer codex gpt-5.4 runtime metadata (openclaw#62694) (thanks @…

78dfce3

…ruclaw7)

obviyus force-pushed the fix/codex-gpt54-resolution branch from e44f4e2 to 78dfce3 Compare April 8, 2026 02:44

obviyus approved these changes Apr 8, 2026

View reviewed changes

obviyus merged commit 81969c7 into openclaw:main Apr 8, 2026
8 checks passed

chatgpt-codex-connector Bot reviewed Apr 8, 2026

View reviewed changes

This was referenced Apr 8, 2026

Update ghcr.io/openclaw/openclaw Docker tag to v2026.4.8 claytono/infra#1923

Merged

📡 Upstream Digest — 2026-04-08 04:30 UTC curtismercier/openclaw-mods#505

Open

clawsweeper Bot mentioned this pull request May 1, 2026

fix: cap effective session context window by known model limit #39859

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Agents: prefer runtime-resolved metadata for explicit codex gpt-5.4#62694

Agents: prefer runtime-resolved metadata for explicit codex gpt-5.4#62694
obviyus merged 6 commits intoopenclaw:mainfrom
ruclaw7:fix/codex-gpt54-resolution

ruclaw7 commented Apr 7, 2026 •

edited

Loading

Uh oh!

greptile-apps Bot commented Apr 7, 2026

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Uh oh!

Uh oh!

ruclaw7 commented Apr 8, 2026

Uh oh!

obviyus left a comment

Uh oh!

Uh oh!

obviyus commented Apr 8, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Uh oh!

chatgpt-codex-connector Bot Apr 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

ruclaw7 commented Apr 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Testing

Uh oh!

greptile-apps Bot commented Apr 7, 2026

Greptile Summary

Confidence Score: 5/5

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

ruclaw7 commented Apr 8, 2026

Uh oh!

obviyus left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

obviyus commented Apr 8, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Apr 8, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ruclaw7 commented Apr 7, 2026 •

edited

Loading