fix(strategy29): balance recent signals and simplify summary diagnostics

[jackhai9]

One-minute signals could fill the recent cross-timeframe list and hide retained signals from longer intervals. Keep the latest three signals for each configured timeframe, then sort the combined list by close time and sequence. Intervals without retained signals show an empty-history explanation instead of fabricated rows.

The main panel focuses on the symbol, connection, monitoring coverage, event check time, signals, and actionable exceptions. Specification and local code versions, universe details, processing rows, and global notification diagnostics move into a collapsed diagnostics section. Processing rows use interval duration order.

Strategy29 requests the bounded latest_per_timeframe snapshot and resets its cursor when the nonempty configured interval set changes. The shared request whitelist is updated in both Strategy29 0.5.5 and its Strategy27 0.6.5 provider. The client requires v4 API envelopes and retains v2 event identities. An old backend is rejected before event requests; a backend version change between status and events cannot publish a stale-version page or advance the cursor. Install identities and update URLs are preserved.

Deployment dependency: https://github.com/jackhai9/CorsairQuant/pull/344 must finish deployment before this PR is merged and the new clients are installed.

Validation:

• Full npm test: 1,032 passed. Playwright UI suite: 83 passed.
• Both affected builds, generated JavaScript checks, release-contract validation, and git diff --check passed.
• Cross-repository verification passed through real SQLite, the Python HTTP gateway, and the generated Strategy29 validator using v4 envelopes and v2 event records.
• Rendered normal, expanded-diagnostics, narrow, data-gap, and disabled-module previews were inspected. Checks cover per-interval quotas, duration order, preserved visible exceptions, English and Chinese copy, panel position, and no page errors.
• Independent read-only reviewer userscript_final_review returned GO on 2026-09-13 with no actionable findings for the summary implementation. The reviewer independently ran 57 focused tests, checked both generated files byte-for-byte using in-memory builds, inspected existing rendered previews, and reported no write actions. The version-boundary follow-up review is recorded below.

Artifact SHA-256:

• Strategy29 0.5.5: dd7d989a64c24b8d2956fa73d52e83bb811687b0f1f0efbfc9b394d3ffe4c427.
• Strategy27 0.6.5: be0db522b84864245c086d30190dd9bf55e245cb879667bd23ee800d90e7c2b8.

Rendered checks use synthetic pages. Production Binance behavior and installed-source verification are post-release checks. Signal detection, notification enablement, and trading actions are unchanged.

Version-boundary follow-up review: api_version_final_review, 2026-09-13, GO with no remaining findings after correcting two current documentation references to V4. The reviewer independently passed 31 client contract tests and verified the generated artifact byte-for-byte using an in-memory build. No write actions were performed; post-review status and reflog matched the primary agent's changes.