fix(agents): skip forced store=true when model declares supportsStore=false by MoerAI · Pull Request #25088 · openclaw/openclaw

MoerAI · 2026-02-24T05:32:55Z

Summary

When using the azure-responses provider with compat.supportsStore=false, the adapter still forces store=true in the request payload. Azure OpenAI Responses API then returns HTTP 400 because it references a prior rs_* response ID that was never persisted.

Root Cause

createOpenAIResponsesStoreWrapper forces store=true for all OpenAI Responses API providers via shouldForceResponsesStore(). This function checks the API type and provider but never checks the compat.supportsStore flag, so models configured with supportsStore: false still get store=true injected.

Fix

Added an early return in shouldForceResponsesStore() that checks model.compat?.supportsStore === false. When this flag is explicitly set to false, the function returns false and the store wrapper skips the store=true injection.

Before

{
  "providers": {
    "azure-responses": {
      "models": [{ "id": "gpt-5.2", "compat": { "supportsStore": false } }]
    }
  }
}

→ HTTP 400: Item with id 'rs_...' not found. Items are not persisted when store is set to false.

After

With supportsStore: false, the adapter no longer injects store=true, preventing the 400 error.

Verification

LSP diagnostics clean
The check is minimal and early-exit — no behavioral change for providers that don't set supportsStore: false

Greptile Summary

Adds an early-exit check in shouldForceResponsesStore() to respect model.compat.supportsStore === false, preventing forced store=true injection for models that explicitly declare they don't support server-side persistence.

The fix is a minimal, well-placed guard that short-circuits before the existing API/provider checks
Uses strict === false comparison, so only explicit opt-outs are affected — no behavioral change for models that omit the flag
The compat type annotation aligns with the existing ModelCompatConfig type and Zod schema (ModelCompatSchema)
Note: the PR description's Azure scenario may not perfectly match the current guard logic (Azure providers typically wouldn't pass the OPENAI_RESPONSES_PROVIDERS check), but the fix is still correct as a general-purpose guard for any provider/API combination where supportsStore: false is set
No tests are included; the shouldForceResponsesStore function had no tests prior to this PR either

Confidence Score: 4/5

This PR is safe to merge — a minimal, defensive early-return guard with no side effects on existing behavior.
The change is small (6 lines), logically correct (strict === false check), properly typed, and doesn't alter behavior for any model that doesn't explicitly set supportsStore: false. No tests are included but the risk surface is very low.
No files require special attention.

_{Last reviewed commit: f4db921}

SamuelHinestrosa

Revisión automática: Cambios revisados. Asegurar que los tests pasen.

…=false (openclaw#25058)

, #22441, #25088) Landed from contributor PRs #16930, #22441, and #25088. Co-authored-by: liweiguang <[email protected]> Co-authored-by: EdwardWu7 <[email protected]> Co-authored-by: MoerAI <[email protected]>

steipete · 2026-02-27T16:16:15Z

Landed on main as 8da3a9a92d9c2c6db6fa6caa881e9310d1792e93.

Rewrite summary:

Kept fix(agents): skip forced store=true when model declares supportsStore=false #25088 behavior: do not force Responses store=true when compat.supportsStore=false.
Unified with fix: treat empty baseUrl as non-direct for OpenAI Responses #16930 + fix: force store for openai-responses providers #22441 into one OpenAI Responses context wrapper.
Added automatic OpenAI server-side context_management compaction injection for compatible direct OpenAI models.
Added tests, docs, and one unified changelog entry.

SHA notes:

Original PR commit: b0fc4350f5a71d774f745cac5cf5189f96050a21
Landed commit: 8da3a9a92d9c2c6db6fa6caa881e9310d1792e93

Thanks @MoerAI.

…nclaw#16930, openclaw#22441, openclaw#25088) Landed from contributor PRs openclaw#16930, openclaw#22441, and openclaw#25088. Co-authored-by: liweiguang <[email protected]> Co-authored-by: EdwardWu7 <[email protected]> Co-authored-by: MoerAI <[email protected]>

…nclaw#16930, openclaw#22441, openclaw#25088) Landed from contributor PRs openclaw#16930, openclaw#22441, and openclaw#25088. Co-authored-by: liweiguang <[email protected]> Co-authored-by: EdwardWu7 <[email protected]> Co-authored-by: MoerAI <[email protected]> (cherry picked from commit c38be8f)

…nclaw#16930, openclaw#22441, openclaw#25088) Landed from contributor PRs openclaw#16930, openclaw#22441, and openclaw#25088. Co-authored-by: liweiguang <[email protected]> Co-authored-by: EdwardWu7 <[email protected]> Co-authored-by: MoerAI <[email protected]>

openclaw-barnacle bot added agents Agent runtime and tooling size: XS labels Feb 24, 2026

SamuelHinestrosa suggested changes Feb 24, 2026

View reviewed changes

LalithaMV approved these changes Feb 24, 2026

View reviewed changes

MoerAI force-pushed the fix/azure-responses-supports-store branch from f4db921 to 956acbe Compare February 24, 2026 07:22

fix(agents): skip forced store=true when model declares supportsStore…

b0fc435

…=false (openclaw#25058)

MoerAI force-pushed the fix/azure-responses-supports-store branch from 956acbe to b0fc435 Compare February 27, 2026 12:11

steipete self-assigned this Feb 27, 2026

This was referenced Feb 27, 2026

fix: treat empty baseUrl as non-direct for OpenAI Responses #16930

Closed

fix: force store for openai-responses providers #22441

Closed

steipete closed this Feb 27, 2026

gemini-code-assist bot mentioned this pull request Mar 1, 2026

chore(sync): replay rollup fixes on upstream/main MillionthOdin16/openclaw#111

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix(agents): skip forced store=true when model declares supportsStore=false#25088

fix(agents): skip forced store=true when model declares supportsStore=false#25088
MoerAI wants to merge 1 commit intoopenclaw:mainfrom
MoerAI:fix/azure-responses-supports-store

MoerAI commented Feb 24, 2026 •

edited by greptile-apps bot

Loading

Uh oh!

SamuelHinestrosa left a comment

Uh oh!

steipete commented Feb 27, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Uh oh!

Conversation

MoerAI commented Feb 24, 2026 • edited by greptile-apps bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Root Cause

Fix

Before

After

Verification

Greptile Summary

Confidence Score: 4/5

Uh oh!

SamuelHinestrosa left a comment

Choose a reason for hiding this comment

Uh oh!

steipete commented Feb 27, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

MoerAI commented Feb 24, 2026 •

edited by greptile-apps bot

Loading