fix: five correctness follow-ups to the catalog metadata work - #2030
Conversation
- Normalize a configured effort value (case/whitespace) before thinking resolution on both engines, so "OFF" is read as off instead of being sent upstream as an invalid effort. - Clamp a declared input cap to the effective context window in the effective model resolution, so an override lowering max_context_size cannot leave a stale larger cap behind. - Report context usage against the same effective cap (max_input_tokens ?? max_context_tokens) in the SDK getStatus and the v2 legacy status projections, matching the session-event and service surfaces. - Preserve concretely declared per-model endpoints when the override npm is unrecognized, via the same OpenAI-compatible fallback used for top-level entries. - Attribute resolved.maxInputSize and resolved.capabilities .max_input_tokens in the model inspector to their config, override, or clamp provenance.
🦋 Changeset detectedLatest commit: 7c3d774 The changes in this PR will be included in the next version bump. This PR includes changesets to release 5 packages
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
commit: |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 1dfad4ceda
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".
…status - The overflow-learned provider window is now written into both max_context_tokens and max_input_tokens of the effective compaction context, so the strategy cannot bypass it by re-selecting the raw catalog input cap during overflow recovery. - agent.status.updated publishes max_input_tokens ?? max_context_tokens, matching the other v1/v2 status surfaces.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 82ffdaf386
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".
…etary override SDKs - effectiveModelAlias / effectiveModelConfig now build a copy before clamping maxInputSize to the effective window instead of rewriting the caller's config record in place. - context_usage is clamped to 1 in the v2 legacy status, the v1 session service, and the SDK getStatus; the default-model status fallback also resolves the input cap. - Overrides naming known proprietary SDKs (Bedrock, Cohere) are refused before the OpenAI-compatible fallback, matching top-level import behavior. - The inspector attributes a clamped maxInputSize to the clamp even when the raw value came from models.*.overrides.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 8faae90434
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".
…w event ratio - Whitespace-only configured efforts now read as absent, and the KIMI_MODEL_THINKING_EFFORT override is lowercased on both engines. - kap-server's WS legacy status publishes max_input_tokens ?? max_context_tokens like the other status surfaces. - SDK getStatus keeps the ratio unclamped on purpose (>100% is the documented overflow signal on that path); schema-bounded REST status stays clamped to 1. - Pin the provider-observed window beating a declared input cap with a v1 full-compaction regression test.
1 participant
Related Issue
Follow-up to #2015 (review comments from the automated review).
Problem
Five small correctness issues remained after the catalog metadata work:
thinking.effort = "OFF"(or with whitespace) was treated as a concrete effort and could be sent upstream as an invalidreasoning_effort— the clamp comparison was case-sensitive and the value was never normalized.getStatusand the v2 legacy status used the total window.@ai-sdk/xai) silently dropped the endpoint and routed the model to the provider default.resolved.maxInputSize/resolved.capabilities.max_input_tokenswithout provenance, breaking the per-value source contract.What changed
maxInputSizeto the effectivemaxContextSizein both effective-model resolvers.max_input_tokens ?? max_context_tokensas the usage denominator (and reported cap) in the SDKgetStatusand the v2 legacy status projection.Each fix carries targeted tests; full suites pass on kosong, node-sdk, agent-core, agent-core-v2, and the CLI, plus typecheck and full-repo oxlint.
Checklist
gen-changesetsskill, or this PR needs no changeset.gen-docsskill, or this PR needs no doc update.