[backryun]

Summary

• Follow up on the maintainer feedback in fix(release): repair build and quality gate drift #8177 with a focused branch against the current release/v3.8.49 tip.
• Sanitize Muse Spark GraphQL fetch exceptions before they reach either executor logs or client-facing error handling.
• Resolve the sql.js WASM asset from the package's public entrypoint and remove the unexported sql.js/package.json lookup.
• Remove the now-obsolete Muse Spark exception from the error-helper gate.

Scope

This intentionally excludes the overlapping base-red reconciliation work from #8177:

• reasoningPlaceholder is owned by fix(resilience,translator): release/v3.8.49 base-red regressions (slice 1) + eslint baseline #8254.
• the adaptive compression ladder is owned by fix(compression): rank codex-responses engine in adaptive ladder maps (base-red slice 2) #8256.
• chat.ts / chatPredicates.ts are owned by fix(sse): export PROVIDER_BREAKER_FAILURE_STATUSES — fix ReferenceError in chat.ts (base-red slice 4) #8258.
• complexity and ESLint baseline reconciliation remain release-captain work.

Only the two runtime fixes that remain unique on the current release tip are included.

Root cause

Muse Spark sanitized its HTTP response after a GraphQL fetch failure, but logged the intermediate error first. That left absolute source paths and stack-tail text visible in executor logs.

The sql.js fallback resolved sql.js/package.json, but sql.js does not export that private subpath. The already-present public sql.js entrypoint resolves beside sql-wasm.wasm, so the metadata lookup is unnecessary and causes repeated Turbopack diagnostics.

Validation

• node --import tsx/esm --test tests/unit/muse-spark-web-continuation.test.ts tests/unit/sqljs-build-warning-8135.test.ts — 8/8 passed.
• npm run typecheck:core — passed.
• node scripts/check/check-error-helper.mjs — passed with zero frozen exceptions.
• npm run build — passed; the sql.js/package.json missing-subpath warning is absent.
• git diff --check — passed.

[diegosouzapw]

Merged into release/v3.8.49 — thanks @backryun! Validated via local merge-train (32-core .113) + a per-file discriminator confirming zero regressions vs the pure release tip; CI reds on this PR were pre-existing base-red drift (tracked separately). 🙏