refactor(javascript): own coverage facts and reuse block prefixes

Separate syntax, semantic uncertainty, examination limits and observed Electron applicability at the static producer, then preserve them through worker, graph and retained/inline summary views. Derive the established parse_status from the same facts and leave historical measurements unknown. Give semantic resource-limit classifications one schema owner.

Replace repeated block-prefix searches and copies with collector-scoped syntax indexes. Keep invocation environments separate, charge indexing/replay to the existing work budget and yield during prefix work. Remove the superseded synchronous collector entrypoint.

Validated producer, worker, CLI/MCP, historical and partial-coverage regressions, plus actual cancellation during indexing. Bounded authored 128/256-call controls preserved all prior facts while replacing repeated search/copy work with one index pass; this is not a throughput or agent-token benchmark.

src/domain/javascript/javascriptReturnedReferenceCoverage.test.ts

@@ -21,6 +21,9 @@ it("preserves possible ambient effects and discloses projection widening in sour
   expect(ir.limitations).toContain(RETURNED_REFERENCE_WIDENING_LIMITATION);
   const facts = analyzeJavaScriptStaticSource(source);
   expect(facts.limitations).toContain(RETURNED_REFERENCE_WIDENING_LIMITATION);
+  expect(facts.analysis_coverage?.semantics.uncertainty).toContain(
+    "returned-reference-widening",
+  );
   expect(facts.endpoints).toEqual(
     expect.arrayContaining([
       expect.objectContaining({ value: "/preserved", mechanism: "call:fetch" }),