Execution: optimize transcript search allocation (PR #897)

[groupthinking]

Parent governance program

Progresses #898. This issue is the single work definition for canonical PR #897.

Outcome

Normalize the transcript search query once per memoized filtering pass instead of once per segment, preserving null/empty-query behavior.

Current state — CODE GREEN, PROVENANCE/DEPLOYMENT EVIDENCE BLOCKED

• Canonical PR: perf(web): hoist transcript search normalization #897
• Canonical branch: bolt-optimize-interactive-transcript-4719841486477851193
• Exact head: aca4199ac6e7367fe7542886e4d25efbc163a31e
• Branch position: current-main based; 0 commits behind at verification
• Scope: InteractiveTranscript.tsx plus the Jules learning note
• Review: approved; Copilot reviewed 2/2 files; 0 unresolved threads
• Duplicate/unsafe ⚡ Bolt: Optimize InteractiveTranscript filtering by moving lowercase outside loop #895: closed
• CI 29791637772, Coverage, CodeQL, Security Scan, Secret Scan, and Dependency Review: passed
• E2E 29791637770: repository-skipped
• Blocking statuses: Vercel failed; agent-completion/truth-gate/pr-897 failed

Acceptance criteria

• Lowercase normalization is hoisted outside the segment loop.
• Null and empty search behavior is preserved.
• Implementation scope is focused and review-complete.
• Unsafe duplicate ⚡ Bolt: Optimize InteractiveTranscript filtering by moving lowercase outside loop #895 is closed without transplanting its workflow changes.
• Repair one-issue agent provenance and obtain a valid exact-head truth-gate result without weakening policy.
• Verify the exact-head Vercel Preview or document a path-based non-applicability rule accepted by repository governance.
• Reverify the unchanged exact head and mark perf(web): hoist transcript search normalization #897 ready only when required statuses pass.

Next executable action

Re-run/evaluate the truth gate after this canonical issue and PR manifest are linked. If the gate correctly rejects post-PR intent creation, keep the PR draft and escalate that security-policy boundary; do not bypass it or create a competing implementation.

Execution receipt

• Agent login: google-labs-jules[bot]
• Run ID: 4719841486477851193
• Canonical PR/branch: perf(web): hoist transcript search normalization #897 / bolt-optimize-interactive-transcript-4719841486477851193
• Original PR claimed: 2026-07-20T17:57:06Z
• Latest verified heartbeat: 2026-07-21T07:55:00Z
• Exact head: aca4199ac6e7367fe7542886e4d25efbc163a31e
• Artifact: focused two-file patch, approved review, green code/security workflows

Safety

No merge or branch deletion while current required statuses fail.

Agent task contract

Agent login

google-labs-jules[bot]

Run ID

4719841486477851193

Objective

Hoist guarded transcript-query normalization out of the per-segment filter loop on canonical PR #897 without changing null or empty-query behavior.

Declared file scope

• apps/web/src/components/InteractiveTranscript.tsx
• .jules/bolt.md

Allowed extra files

None.

Focused test paths

• apps/web/src/components/tests/InteractiveTranscript.test.tsx

Unrestricted scope

• Unrestricted scope is not requested.

Latest truth-gate result

Workflow run: 29812642843

• Identity projection now succeeds for issue Execution: optimize transcript search allocation (PR #897) #908, google-labs-jules[bot], run 4719841486477851193.
• Structural blocker: missing_intent_snapshot because the canonical GitHub issue was created after the PR.
• Other current reasons: missing trusted agent result, missing current-head Copilot policy evidence, draft state, and missing focused-test evidence.
• The invalid PR title was corrected after this run.

This is a provenance/security-policy boundary, not a reason to weaken or bypass the gate. Keep the PR draft while an authorized owner chooses whether to retire the old PR or establish a policy-compliant replacement after a trusted intent snapshot.

[linear-code]

MYX-90

[github-actions]

👋 Thanks for opening this issue! It has been automatically labeled for triage.

[coderabbitai]

🔗 Related PRs

#266 - Add transcription-driven action agent (tool/action layer) [merged]
#277 - Vercel remediation complete: waitUntil adoption + active middleware + @vercel/functions package (resolves three user-defined outcomes; 16-agent verification-gate consolidated PASS on full matrix) [closed]
#404 - fix(web): sync package-lock with @vercel/functions (unblocks #277 auto-merge) [merged]
#648 - fix: address CodeRabbit review findings on merged perf pass (#642) [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!