refactor(eval)!: remove batch insights and online insights

[jariy17]

Description

What: Removes batch insights and online insights from the CLI.

• agentcore eval batch-insights run|get|list|stop
• agentcore eval online-insight create|get|list|update|pause|resume|delete
• agentcore add online-insight / agentcore remove online-insight

Why: Standalone insights will be removed for now.
How: Removed all handlers, core clients and tuis that are part of those.

Link: https://github.com/aws/agentcore-l3-cdk-constructs/pull/420

[agentcore-devx-automation]

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

[agentcore-devx-automation]

AgentCore Harness Review

Verdict: Looks good

Reviewed this as a mechanical removal of the batch-insights and online-insight surfaces on the unreleased refactor branch. The deletions are consistent across the stack:

• Routes and screen imports in src/components/Root.tsx are removed in lockstep with the deleted screens.
• createOnlineInsightHandler / createBatchInsightsHandler are removed from src/handlers/eval/index.tsx and the add/remove routers; corresponding entries in supportedTuiCommands, AddResourceInput, RemoveResourceInput, RemovableResourceType, and ChinaAddKind (via the union) all fall out cleanly.
• EvalClient and TestEvalClient lose the insight methods, and the inputs/responses are removed from src/handlers/eval/types.tsx; FilteredPaginator is still used by src/core/gateway.tsx, so the import removal in src/core/eval.tsx is correct.
• The schema change in src/projectSchemas/online-eval-config.ts (dropping insights/clusteringConfig, making evaluators required with min(1)) is matched by the test updates and by createAddOnlineEvalHandler marking --evaluators required. Existing spec files that only had insights will now fail validation, which is the stated intent for the unreleased branch.
• No stale references to online-insight / batch-insights / clusteringConfig remain in src/ (the only hits are literal substrings inside unrelated batch-evaluation fixture names, and the surviving CloudWatch Logs Insights path called out in the PR body).

No telemetry concerns (feature removal) and no new mocking patterns introduced. Safe to merge.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.52%. Comparing base (5d81751) to head (4a2a4aa).

Additional details and impacted files

@@             Coverage Diff              @@
##           refactor    #2576      +/-   ##
============================================
- Coverage     97.54%   97.52%   -0.02%     
============================================
  Files           657      634      -23     
  Lines         48779    47260    -1519     
============================================
- Hits          47580    46090    -1490     
+ Misses         1199     1170      -29     

☔ 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.

[nborges-aws]

LGTM. One nit, command.md needs to be updated as well to remove the insights commands and flip evaluators to be required