Conversation
mydea
added this pull request to stack #24359
September 14, 2026 08:12
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Want reviews to match your repository better? Bugbot Learning can learn team-specific rules from PR activity. A team admin can enable Learning in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 58c1aaa. Configure here.
| }) | ||
| .start() | ||
| .completed(); | ||
| }); |
There was a problem hiding this comment.
Cache token test not skipped on v6
Medium Severity
The new cache-token test is registered for both matrix versions, but v6 cannot derive cache counts from this gateway scenario. Nearby version-specific cases use test.skipIf, so the v6 entry will assert gen_ai.usage.input_tokens.cache_read and cache_creation values that that path never writes.
Triggered by project rule: PR Review Guidelines for Cursor Bot
Reviewed by Cursor Bugbot for commit 58c1aaa. Configure here.
Member
Author
There was a problem hiding this comment.
this seems to work just fine
Contributor
size-limit report 📦
|
mydea
marked this pull request as ready for review
September 14, 2026 08:48
mydea
requested review from
JPeer264,
RulaKhaled,
isaacs and
sergical
and removed request for
a team
September 14, 2026 08:48
![sentry[bot]](https://avatars.githubusercontent.com/in/12637?s=60&v=4){kind=small}
Comment on lines
+833
to
+843
| const invokeAgent = container.items.find( | ||
| span => span.attributes['sentry.op']?.value === 'gen_ai.invoke_agent', | ||
| )!; | ||
| expect(invokeAgent).toBeDefined(); | ||
| expect(invokeAgent.attributes[GEN_AI_USAGE_CACHE_READ_INPUT_TOKENS]?.value).toBe(80); | ||
| expect(invokeAgent.attributes[GEN_AI_USAGE_CACHE_CREATION_INPUT_TOKENS]?.value).toBe(20); | ||
| }, | ||
| }) | ||
| .start() | ||
| .completed(); | ||
| }); |
Contributor
There was a problem hiding this comment.
Bug: The test 'reads cache token counts from the SDK usage object' is missing a test.skipIf(version === '6') condition, causing it to fail for the v6 test matrix.
Severity: LOW
Suggested Fix
Wrap the test case 'reads cache token counts from the SDK usage object' with a skip condition to prevent it from running against v6. Use test.skipIf(version === '6') around the test definition, similar to other version-specific tests in the suite.
Prompt for AI Agent
Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.
Location:
dev-packages/node-integration-tests/suites/tracing/vercelai/v6_v7/test.ts#L820-L843
Potential issue: The integration test `'reads cache token counts from the SDK usage
object'` is intended to run for both v6 and v7 of the Vercel AI SDK. However, the test
is missing a required skip condition for v6. The test asserts that cache token counts
are derived from the SDK's usage object, a feature specific to v7. The v6 OTel
integration derives these counts from `providerMetadata` and cannot handle the `gateway`
provider metadata shape used in this test scenario. As a result, the test will
incorrectly execute and fail during CI runs for v6.
Did we get this right? 👍 / 👎 to inform future reviews.
JPeer264
approved these changes
Sep 14, 2026
Add a node integration test that runs the real `ai@^7` package through the channel subscriber with `providerMetadata` keyed `gateway` (the AI Gateway shape), asserting cache read/write counts are still derived from the SDK's normalized usage. Skipped on v6, whose OTel processor only reads `providerMetadata`. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
mydea
force-pushed
the
feat/vercel-ai-cache-tokens-node-test
branch
from
September 14, 2026 09:39
58c1aaa to
27897d2
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Stacked on #24350. Adds an end-to-end node integration test that exercises the cache-token fix against the real
ai@^7package, not just the unit test in the base PR.The scenario drives
generateTextthrough a mock model whose usage carries cache read/write counts in the v7 shape (inputTokens.{cacheRead,cacheWrite}) whileproviderMetadatais keyedgateway— the Vercel AI Gateway shape that has no provider-specific cache key. Without the base PR the channel subscriber drops these counts; the test asserts they land on both thegenerate_contentandinvoke_agentspans. It's skipped on v6, whose OTel-processor path derives cache counts fromproviderMetadataalone and can't see the gateway case.🤖 Generated with Claude Code