fix(sec): seal merger-proxy + redemption + transactional SPAC deal recompute

src/sec/forms/miscellaneous-filings/redemption8k.injection.test.ts

@@ -0,0 +1,126 @@
+/**
+ * @license
+ * Copyright 2026 Steven Roussey <sroussey@gmail.com>
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { afterEach, beforeEach, describe, expect, it } from "bun:test";
+import { resetDependencyInjectionsForTesting } from "../../../config/TestingDI";
+import { setupAllDatabases } from "../../../config/setupAllDatabases";
+import { SpacReportWriter } from "../../../storage/spac/SpacReportWriter";
+import { SpacRedemptionExtractionRepo } from "../../../storage/spac/SpacRedemptionExtractionRepo";
+import { ExtractionDeadLetterRepo } from "../../../storage/dead-letter/ExtractionDeadLetterRepo";
+import { MAX_STORED_SPAN_CHARS } from "../registration-statements/s1/verifySourceSpan";
+import {
+  fakeS1Model,
+  registerFakeStructuredProvider,
+} from "../registration-statements/s1/testing/fakeStructuredProvider";
+import { processRedemption8K } from "./redemption8k";
+
+const FULL_TXT =
+  "<SEC-HEADER>\nACCESSION NUMBER: 0000000000-26-injection\n</SEC-HEADER>\n" +
+  "<DOCUMENT>\n<TYPE>8-K\n<SEQUENCE>1\n<TEXT>\n<p>Vote results.</p>\n</TEXT>\n</DOCUMENT>\n" +
+  "<DOCUMENT>\n<TYPE>EX-99.1\n<SEQUENCE>2\n<TEXT>\n" +
+  "<p>Holders of 1,234,567 shares elected to redeem for $12,400,000.</p>\n" +
+  "</TEXT>\n</DOCUMENT>\n";
+
+async function seedSpacWithDeal(cik: number): Promise<void> {
+  const writer = new SpacReportWriter();
+  await writer.recordRegistration({
+    cik,
+    accession_number: `${cik}-reg`,
+    filing_date: "2025-12-01",
+    form: "S-1",
+    primary_document: "s1.htm",
+    spac_name: "Redeem SPAC Inc.",
+    spac_sic: 6770,
+  });
+  await writer.recordDealMilestones({
+    cik,
+    accession_number: `${cik}-da`,
+    filing_date: "2026-01-10",
+    form: "8-K",
+    primary_document: null,
+    events: [{ event_type: "definitive_agreement", event_date: "2026-01-10" }],
+  });
+}
+
+describe("processRedemption8K prompt-injection seal", () => {
+  let cleanup: (() => void) | undefined;
+
+  beforeEach(async () => {
+    resetDependencyInjectionsForTesting();
+    await setupAllDatabases();
+  });
+  afterEach(() => {
+    cleanup?.();
+    cleanup = undefined;
+  });
+
+  it("verifyRowSpan rejects a 1001-char source_span at the gate and dead-letters UNVERIFIED_SOURCE_SPAN", async () => {
+    await seedSpacWithDeal(800);
+    const oversizedSpan = "X".repeat(MAX_STORED_SPAN_CHARS + 1);
+    expect(oversizedSpan.length).toBe(1001);
+    const { unregister } = registerFakeStructuredProvider([
+      {
+        redemption_shares: 999,
+        redemption_amount: 999,
+        price_per_share: 1,
+        confidence: 0.99,
+        source_span: oversizedSpan,
+      },
+    ]);
+    cleanup = unregister;
+
+    await processRedemption8K({
+      cik: 800,
+      accession_number: "0000000000-26-injection",
+      filing_date: "2026-03-20",
+      form: "8-K",
+      itemCodes: ["5.07"],
+      fullSubmissionText: FULL_TXT,
+      model: fakeS1Model(),
+    });
+
+    expect(
+      await new SpacRedemptionExtractionRepo().getByAccession("0000000000-26-injection")
+    ).toBeUndefined();
+    const dl = await new ExtractionDeadLetterRepo().listPending("redemption");
+    const red = dl.find((d) => d.section_name === "redemption");
+    expect(red?.reason_code).toBe("UNVERIFIED_SOURCE_SPAN");
+  });
+
+  it("persist site caps the stored source_span via boundSourceSpan at MAX_STORED_SPAN_CHARS", async () => {
+    await seedSpacWithDeal(801);
+    // A verbatim span from the EX-99.1 narrative persists unchanged.
+    const verbatim = "1,234,567 shares elected to redeem for $12,400,000";
+    expect(verbatim.length).toBeLessThanOrEqual(MAX_STORED_SPAN_CHARS);
+    const { unregister } = registerFakeStructuredProvider([
+      {
+        redemption_shares: 1234567,
+        redemption_amount: 12400000,
+        price_per_share: 10.05,
+        confidence: 0.95,
+        source_span: verbatim,
+      },
+    ]);
+    cleanup = unregister;
+
+    await processRedemption8K({
+      cik: 801,
+      accession_number: "0000000000-26-injection-2",
+      filing_date: "2026-03-20",
+      form: "8-K",
+      itemCodes: ["5.07"],
+      fullSubmissionText: FULL_TXT,
+      model: fakeS1Model(),
+    });
+
+    const ext = await new SpacRedemptionExtractionRepo().getByAccession(
+      "0000000000-26-injection-2"
+    );
+    expect(ext).toBeDefined();
+    expect((ext?.source_span ?? "").length).toBeLessThanOrEqual(MAX_STORED_SPAN_CHARS);
+    expect(ext?.source_span).toBe(verbatim);
+  });
+});

src/sec/forms/miscellaneous-filings/redemption8k.ts

@@ -8,7 +8,7 @@ import { globalServiceRegistry, renderMarkdown } from "workglow";
 import { parseEdgarHtml } from "../../html/parseEdgarHtml";
 import { parseEightKSubmission } from "../registration-statements/s1/parseSubmission";
 import { makeRunSection } from "../registration-statements/s1/sectionRunner";
-import { spanAppearsIn } from "../registration-statements/s1/verifySourceSpan";
+import { boundSourceSpan, verifyRowSpan } from "../registration-statements/s1/verifySourceSpan";
 import { extractRedemption } from "../registration-statements/s1/sectionExtractors";
 import type { RedemptionRow } from "../registration-statements/s1/redemptionSchema";
 import {
@@ -113,7 +113,7 @@ export async function processRedemption8K(args: ProcessRedemption8KArgs): Promis
     notFoundDetail: "no primary/EX-99 narrative text",
     emptyDetail: "no redemption returned",
     lowConfidenceDetail: "below confidence floor",
-    verifyRow: (t, r) => spanAppearsIn(t, r.source_span),
+    verifyRow: (t, r) => verifyRowSpan(t, r.source_span),
     unverifiedAllDetail: "redemption source_span not present in narrative text",
     extract: async (t) => {
       const row = await extractRedemption(t, model);
@@ -132,7 +132,7 @@ export async function processRedemption8K(args: ProcessRedemption8KArgs): Promis
         redemption_amount: row.redemption_amount,
         price_per_share: row.price_per_share,
         confidence: row.confidence,
-        source_span: row.source_span,
+        source_span: boundSourceSpan(row.source_span),
         model_id,
         created_at: new Date().toISOString(),
       });

src/sec/forms/proxies-information-statements/Form_DEFM14A.injection.test.ts

@@ -0,0 +1,147 @@
+/**
+ * @license
+ * Copyright 2026 Steven Roussey <sroussey@gmail.com>
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { afterEach, beforeEach, describe, expect, it } from "bun:test";
+import { resetDependencyInjectionsForTesting } from "../../../config/TestingDI";
+import { setupAllDatabases } from "../../../config/setupAllDatabases";
+import { SpacRepo } from "../../../storage/spac/SpacRepo";
+import { SpacReportWriter } from "../../../storage/spac/SpacReportWriter";
+import { SpacMergerExtractionRepo } from "../../../storage/spac/SpacMergerExtractionRepo";
+import { ExtractionDeadLetterRepo } from "../../../storage/dead-letter/ExtractionDeadLetterRepo";
+import { MAX_STORED_SPAN_CHARS } from "../registration-statements/s1/verifySourceSpan";
+import {
+  fakeS1Model,
+  registerFakeStructuredProvider,
+} from "../registration-statements/s1/testing/fakeStructuredProvider";
+import { Form_DEFM14A } from "./Form_DEFM14A";
+import { processMergerProxy } from "./Form_DEFM14A.storage";
+
+const FIXTURE = `${import.meta.dir}/mock_data/merger-proxy/defm14a_sample.txt`;
+
+async function seedSpac(cik: number): Promise<void> {
+  const writer = new SpacReportWriter();
+  await writer.recordRegistration({
+    cik,
+    accession_number: `${cik}-reg`,
+    filing_date: "2020-12-01",
+    form: "S-1",
+    primary_document: "s1.htm",
+    spac_name: "Merge SPAC Inc.",
+    spac_sic: 6770,
+  });
+  await writer.recordDealMilestones({
+    cik,
+    accession_number: `${cik}-da`,
+    filing_date: "2021-03-05",
+    form: "8-K",
+    primary_document: null,
+    events: [{ event_type: "definitive_agreement", event_date: "2021-03-01" }],
+  });
+}
+
+async function runProxy(cik: number, accession_number: string): Promise<void> {
+  const txt = await Bun.file(FIXTURE).text();
+  const parsed = await Form_DEFM14A.parse("DEFM14A", txt);
+  await processMergerProxy({
+    cik,
+    file_number: "",
+    accession_number,
+    filing_date: "2021-05-01",
+    primary_doc: "proxy.htm",
+    form: "DEFM14A",
+    formMergerProxy: parsed,
+    model: fakeS1Model(),
+  });
+}
+
+describe("processMergerProxy prompt-injection seal", () => {
+  let cleanup: (() => void) | undefined;
+
+  beforeEach(async () => {
+    resetDependencyInjectionsForTesting();
+    await setupAllDatabases();
+  });
+  afterEach(() => {
+    cleanup?.();
+    cleanup = undefined;
+  });
+
+  it("verifyRowSpan rejects a 1001-char source_span at the gate, dead-letters UNVERIFIED_SOURCE_SPAN, persists nothing", async () => {
+    await seedSpac(700);
+    // The raw source_span exceeds the storage cap. Even though it would
+    // appear verbatim in a synthetically-large section text, verifyRowSpan
+    // rejects it BEFORE normalization, mirroring the S-1 storage-side cap.
+    const oversizedSpan = "X".repeat(MAX_STORED_SPAN_CHARS + 1);
+    expect(oversizedSpan.length).toBe(1001);
+    const { unregister } = registerFakeStructuredProvider([
+      {
+        target_name: "Mallory Inc.",
+        pipe_amount: 999_999,
+        merger_consideration: "fabricated",
+        confidence: 0.99,
+        source_span: oversizedSpan,
+      },
+    ]);
+    cleanup = unregister;
+
+    await runProxy(700, "700-defm");
+
+    expect(await new SpacMergerExtractionRepo().getByAccession("700-defm")).toBeUndefined();
+    const dl = await new ExtractionDeadLetterRepo().listPending("merger-proxy");
+    const merger = dl.find((d) => d.section_name === "merger");
+    expect(merger?.reason_code).toBe("UNVERIFIED_SOURCE_SPAN");
+  });
+
+  it("persist site caps the stored source_span via boundSourceSpan at MAX_STORED_SPAN_CHARS", async () => {
+    await seedSpac(701);
+    // A row whose source_span verifies (short, present in fixture) persists
+    // unchanged: boundSourceSpan returns the span as-is at-or-below the cap.
+    const verbatim = "business combination with Acme Target Inc.";
+    expect(verbatim.length).toBeLessThanOrEqual(MAX_STORED_SPAN_CHARS);
+    const { unregister } = registerFakeStructuredProvider([
+      {
+        target_name: "Acme Target Inc.",
+        pipe_amount: 150_000_000,
+        merger_consideration: "$10 per share",
+        confidence: 0.95,
+        source_span: verbatim,
+      },
+    ]);
+    cleanup = unregister;
+
+    await runProxy(701, "701-defm");
+
+    const ext = await new SpacMergerExtractionRepo().getByAccession("701-defm");
+    expect(ext).toBeDefined();
+    // Persisted span is bounded at the storage cap (here unchanged, since the
+    // raw span is well below the cap). The contract under test is that the
+    // call site flows through boundSourceSpan rather than persisting the
+    // model output verbatim.
+    expect((ext?.source_span ?? "").length).toBeLessThanOrEqual(MAX_STORED_SPAN_CHARS);
+    expect(ext?.source_span).toBe(verbatim);
+  });
+
+  it("rolling-up after a verifier reject does not surface the rejected target onto the spac row", async () => {
+    await seedSpac(702);
+    const oversizedSpan = "Y".repeat(MAX_STORED_SPAN_CHARS + 1);
+    const { unregister } = registerFakeStructuredProvider([
+      {
+        target_name: "Mallory Inc.",
+        pipe_amount: 1,
+        merger_consideration: "fabricated",
+        confidence: 0.99,
+        source_span: oversizedSpan,
+      },
+    ]);
+    cleanup = unregister;
+
+    await runProxy(702, "702-defm");
+
+    const row = await new SpacRepo().getSpac(702);
+    expect(row?.target_name ?? null).toBeNull();
+    expect(row?.pipe_amount ?? null).toBeNull();
+  });
+});