feat(eval): support session trace filters and log group prefixes

[jariy17]

Summary

• update the AgentCore data-plane SDK to a version that supports CloudWatchFilterConfig.sessionTraceIds
• document sessionTraceIds and logGroupNamePrefixes in the existing batch-evaluation --data-source-config help
• pass raw prefix selectors and session/trace filters through StartBatchEvaluationCommand unchanged
• scope the auto-generated online-eval execution role to logGroupNamePrefixes as well as exact log group names
• cover both workflows through command-handler fixture tests

Testing

• bun test src/handlers/eval/batch-evaluation/batch-evaluation.fixture.test.tsx src/handlers/eval/online-eval/online-eval.test.tsx
• bun test src/core/onlineEvalExecutionRole.test.ts src/handlers/eval/batch-evaluation/batch-evaluation.test.tsx src/handlers/eval/online-eval/online-eval.screen.test.tsx
• bun run typecheck
• bunx oxlint src/core/eval.tsx src/core/onlineEvalExecutionRole.tsx src/handlers/eval/sessionSource.tsx src/handlers/eval/batch-evaluation/batch-evaluation.fixture.test.tsx src/handlers/eval/online-eval/online-eval.test.tsx
• git diff --check

[agentcore-devx-automation]

Claude Security Review: no high-confidence findings. (run)

[agentcore-devx-automation]

AgentCore Harness Review

Verdict: Looks good

Small, well-scoped change: SDK bump to pick up sessionTraceIds, help-text update, and a test covering the raw pass-through path.

• Verified SessionTraceIds / CloudWatchFilterConfig.sessionTraceIds exist in @aws-sdk/client-bedrock-agentcore@3.1136.0, so the ^3.1135.0 bump is sufficient.
• The test in src/core/batchEvaluation.test.tsx mocks at the AWS SDK client boundary (send), which is the right layer — not over-mocked.
• No new user-facing commands are introduced, so no additional telemetry instrumentation is warranted.

One minor thing to double-check (not a blocker): the help text in src/handlers/eval/sessionSource.tsx states traceIds accepts "up to 100 trace IDs". The SDK typings don't encode that limit; worth confirming against the public API docs before this misleads users if the real cap differs. Fine to address in a follow-up.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.26%. Comparing base (a23e472) to head (576581f).

Additional details and impacted files

@@            Coverage Diff            @@
##           refactor    #2345   +/-   ##
=========================================
  Coverage     97.26%   97.26%           
=========================================
  Files           610      610           
  Lines         40590    40605   +15     
=========================================
+ Hits          39480    39495   +15     
  Misses         1110     1110           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[agentcore-devx-automation]

Claude Security Review: no high-confidence findings. (run)