Cap the summary and highlight paths, closing the last ENG-3697 review findings

[Dithilli]

Follow-up to #23, closing the three review findings that were still open when it merged.

Cap summary and highlights (Entelligence MAJOR, simplify.ts:112)

matchedText returned summary verbatim. Summary is the server's concatenation of a hit's highlights, and the API caps each highlight at ~2000 — not their sum — so a hit with ten of them emitted ~20KB. That means Search and Answer, the two most-used operations, were never bounded: #23's table claiming them "bounded, correct" was wrong.

Every path in matchedText is now capped at MAX_SIMPLIFIED_TEXT.

This changes Search and Answer output for existing workflows, so the release that carries it is 0.8.0, not a patch.

Two review nits

• connection-id.test.mjs asserted only that some preSend was wired, which an unrelated no-op would satisfy. Now asserts identity.
• The live-list cap had no test containing any text — the prod fixture has none, so an unbounded payload would have passed. That fixture stays a verbatim recording; the new test brings its own long node.

Verification

npm run build, npm run lint, npm test — 112/112 (was 108).

The summary path is covered by unit tests; end-to-end confirmation through a live query is blocked locally on an expired AWS SSO token (the query path needs SageMaker embeddings), not on anything in this diff.

---

^{Need help on this PR? Tag @codesmith-bot with what you need. Autofix is disabled.}

[linear-code]

ENG-3697

[entelligence-ai-pr-reviews]

EntelligenceAI PR Summary

Caps simplified document text across summary, highlight, and tree-fallback paths to prevent Search and Answer results from expanding to tens of kilobytes. Adds regression coverage for summary/highlight aggregation and live-list payloads, and strengthens the connection ID wiring test to verify the exact validator function.

---

Review Scorecard

Dimension	Rating	Basis
Code Quality	●●●●● 5/5 — Excellent	no critical/significant findings — rated Excellent
Blast Radius	Low	changed symbols are referenced only within their own file(s); no high-impact surface touched, 1 file(s) / ~11 line(s) changed (size only — not a blast signal)
Merge Confidence	●●●●● 5/5 — Safe to Merge	code quality 5/5 × Low blast radius

Safe to merge — this is a clean, low-blast-radius change with no correctness, security, or robustness issues identified. The caps are applied across the summary, highlight, and tree-fallback paths, with regression coverage for aggregation and live-list payloads; the connection ID test also verifies the exact validator function. No pre-existing unresolved issues or new review findings require follow-up.

Key Findings:

• Summary, highlight, and tree-fallback text paths are capped consistently, preventing oversized Search and Answer payloads without changing unrelated behavior.
• Regression coverage exercises summary/highlight aggregation and live-list payloads, while the connection ID wiring test checks the precise validator rather than only broad behavior.

Evaluated against

• 4/4 changed files reviewed
• criteria: correctness, security & access control, robustness & error handling, concurrency & data integrity, repo conventions / steering docs
• steering docs: none found in repo