fix(python): validate shared pyc header facts

src/application/python/PythonPycService.test.ts

@@ -66,6 +66,28 @@ const report: PythonPyc = {
   ],
 };
 
+it.each([
+  {
+    flags: 1,
+    hash_based: true,
+    source_hash_hex: "",
+    source_mtime: null,
+    source_size: null,
+  },
+  { flags: 1 },
+])("refuses inconsistent provider header facts %j", async (header) => {
+  const result = await new PythonPycService({
+    close: async () => {},
+    identity,
+    inspect: async () =>
+      ok({ ...report, header: { ...report.header, ...header } }),
+  }).inspect({ path: report.artifact.path });
+  expect(result).toMatchObject({
+    ok: false,
+    error: { _tag: "AnalysisOutputError" },
+  });
+});
+
 it("rejects a relative path and extra arguments before reading", async () => {
   let calls = 0;
   const service = new PythonPycService({

src/domain/python/pythonPyc.test.ts

@@ -3,9 +3,49 @@ import {
   inspectPythonPycInputSchema,
   pythonPycMarshalProfile,
   PYTHON_PYC_MARSHAL_PROFILES,
+  pythonPycHeaderSchema,
 } from "./pythonPyc.js";
 import { pythonBytecodeIdentity } from "./pythonBytecodeIdentity.js";
 
+const hashHeader = {
+  magic: 3531,
+  python_version: "3.12",
+  code_layout: "py311",
+  flags: 3,
+  hash_based: true,
+  check_source: true,
+  source_hash_hex: "0123456789abcdef",
+  source_mtime: null,
+  source_size: null,
+};
+
+it.each(["", "ab", "g".repeat(16)])(
+  "rejects an invalid source hash %j",
+  (source_hash_hex) => {
+    expect(
+      pythonPycHeaderSchema.safeParse({ ...hashHeader, source_hash_hex })
+        .success,
+    ).toBe(false);
+  },
+);
+
+it.each([
+  { hash_based: false },
+  { check_source: false },
+  { source_hash_hex: null },
+  { source_mtime: 0 },
+  { source_size: 0 },
+])("keeps header facts consistent with the recorded flags: %j", (changed) => {
+  expect(
+    pythonPycHeaderSchema.safeParse({ ...hashHeader, ...changed }).success,
+  ).toBe(false);
+});
+
+it("preserves unknown raw header flags and valid hash bytes", () => {
+  const header = { ...hashHeader, flags: 0x80000003 };
+  expect(pythonPycHeaderSchema.parse(header)).toEqual(header);
+});
+
 it("keeps pinned CPython identity separate from standalone marshal coverage", () => {
   expect(pythonPycMarshalProfile(3394)).toEqual({
     code_layout: "py37",

src/domain/python/pythonPyc.ts

@@ -205,17 +205,53 @@ const pythonCodeObjectSchema = z.strictObject({
 });
 export type PythonPycCodeObject = z.infer<typeof pythonCodeObjectSchema>;
 
-export const pythonPycHeaderSchema = z.strictObject({
-  magic: z.number().int().min(0).max(0xffff),
-  python_version: z.string().min(1),
-  code_layout: z.enum(PYTHON_PYC_CODE_LAYOUTS),
-  flags: uint,
-  hash_based: z.boolean(),
-  check_source: z.boolean(),
-  source_hash_hex: z.string().nullable(),
-  source_mtime: uint.nullable(),
-  source_size: uint.nullable(),
-});
+export const pythonPycHeaderSchema = z
+  .strictObject({
+    magic: z.number().int().min(0).max(0xffff),
+    python_version: z.string().min(1),
+    code_layout: z.enum(PYTHON_PYC_CODE_LAYOUTS),
+    flags: uint,
+    hash_based: z.boolean(),
+    check_source: z.boolean(),
+    source_hash_hex: z
+      .string()
+      .regex(/^[0-9a-f]{16}$/u)
+      .nullable(),
+    source_mtime: uint.nullable(),
+    source_size: uint.nullable(),
+  })
+  .superRefine((header, context) => {
+    for (const [field, flag] of [
+      ["hash_based", PYTHON_PYC_HASH_BASED],
+      ["check_source", PYTHON_PYC_CHECK_SOURCE],
+    ] as const)
+      if (header[field] !== ((header.flags & flag) !== 0))
+        context.addIssue({
+          code: "custom",
+          path: [field],
+          message: `${field} disagrees with its recorded header flag.`,
+        });
+    if (header.hash_based === (header.source_hash_hex === null))
+      context.addIssue({
+        code: "custom",
+        path: ["source_hash_hex"],
+        message:
+          "Hash-based pycs require an 8-byte source hash; others must omit it.",
+      });
+    if (header.hash_based === (header.source_mtime !== null))
+      context.addIssue({
+        code: "custom",
+        path: ["source_mtime"],
+        message: "Timestamp pycs require mtime; hash-based pycs must omit it.",
+      });
+    if (header.hash_based === (header.source_size !== null))
+      context.addIssue({
+        code: "custom",
+        path: ["source_size"],
+        message:
+          "Timestamp pycs require source_size; hash-based pycs must omit it.",
+      });
+  });
 
 /** Bounded CPython 3.7–3.13 pyc observation; marshal is walked, bytecode is not decompiled. */
 export const pythonPycSchema = z
@@ -253,29 +289,6 @@ export const pythonPycSchema = z
         message: "Marshal byte count disagrees with the header-stripped file.",
       });
     }
-    if (report.header.hash_based === (report.header.source_hash_hex === null)) {
-      context.addIssue({
-        code: "custom",
-        path: ["header", "source_hash_hex"],
-        message:
-          "Hash-based pycs require an 8-byte source hash; others must omit it.",
-      });
-    }
-    if (report.header.hash_based === (report.header.source_mtime !== null)) {
-      context.addIssue({
-        code: "custom",
-        path: ["header", "source_mtime"],
-        message: "Timestamp pycs require mtime; hash-based pycs must omit it.",
-      });
-    }
-    if (report.header.hash_based === (report.header.source_size !== null)) {
-      context.addIssue({
-        code: "custom",
-        path: ["header", "source_size"],
-        message:
-          "Timestamp pycs require source_size; hash-based pycs must omit it.",
-      });
-    }
     if (report.coverage.code_objects !== report.code_objects.length) {
       context.addIssue({
         code: "custom",

src/domain/windows/ntfsMft.test.ts

@@ -127,17 +127,10 @@ describe("MFT header facts", () => {
     });
   });
 
-  it.each([
-    [0, 1],
-    [49, 1],
-    [64, 1],
-    [72, 1],
-    [48, 1],
-    [48, 0],
-  ])(
-    "reports invalid first attribute offset %i with USA sequence %i",
-    (offset, sequenceNumber) => {
-      const record = recordOf({ sequenceNumber });
+  it.each([0, 49, 64, 72])(
+    "reports an invalid first attribute offset %i",
+    (offset) => {
+      const record = recordOf();
       new DataView(record.buffer).setUint16(20, offset, true);
       const parsed = parseMftImage(record);
       if (!parsed.ok) throw new Error(parsed.error.message);
@@ -146,29 +139,20 @@ describe("MFT header facts", () => {
     },
   );
 
-  it.each([1, 0, 0xffff])(
-    "reports a contradictory declared record number with USA sequence %i",
-    (sequenceNumber) => {
-      const bytes = streamOf(recordOf(), recordOf({ sequenceNumber }));
-      new DataView(bytes.buffer).setUint32(1024 + 44, 0, true);
-      const parsed = parseMftImage(bytes);
-      if (!parsed.ok) throw new Error(parsed.error.message);
-      expect(parsed.value.records[1]).toMatchObject({
-        index: 1,
-        declaredRecordNumber: 0,
-        recordNumberMatchesSlot: false,
-        fixups: {
-          malformed: sequenceNumber === 0 || sequenceNumber === 0xffff,
-        },
-      });
-      expect(parsed.value.metadata.fixup_unverified_record_count).toBe(
-        sequenceNumber === 1 ? 0 : 1,
-      );
-      expect(parsed.value.limitations.join(" ")).toContain(
-        "disagrees with their stream slot",
-      );
-    },
-  );
+  it("reports a declared record number that contradicts its stream position", () => {
+    const bytes = streamOf(recordOf(), recordOf());
+    new DataView(bytes.buffer).setUint32(1024 + 44, 0, true);
+    const parsed = parseMftImage(bytes);
+    if (!parsed.ok) throw new Error(parsed.error.message);
+    expect(parsed.value.records[1]).toMatchObject({
+      index: 1,
+      declaredRecordNumber: 0,
+      recordNumberMatchesSlot: false,
+    });
+    expect(parsed.value.limitations.join(" ")).toContain(
+      "disagrees with their stream slot",
+    );
+  });
 
   it("keeps the inline header window separate from the parsed record count", () => {
     const parsed = parseMftImage(

src/domain/windows/ntfsMft.ts

@@ -373,13 +373,8 @@ const readRecord = (
     true,
   );
 
-  const usaLayoutValid = updateSequenceArrayFits(
-    usaOffset,
-    fixups.entries,
-    recordBytes / FIXUP_SECTOR_BYTES,
-  );
   const declaredRecordNumber =
-    usaLayoutValid && usaOffset >= RECORD_HEADER_WITH_NUMBER_BYTES
+    !fixups.malformed && usaOffset >= RECORD_HEADER_WITH_NUMBER_BYTES
       ? view.getUint32(RECORD_HEADER.recordNumber, true)
       : null;
 
@@ -401,7 +396,7 @@ const readRecord = (
       firstAttributeOffset % 8 === 0 &&
       firstAttributeOffset >= NTFS_MFT_MIN_RECORD_BYTES &&
       firstAttributeOffset + 4 <= usedBytes &&
-      (!usaLayoutValid ||
+      (fixups.malformed ||
         firstAttributeOffset >= usaOffset + fixups.entries * 2),
     baseRecordReference:
       baseRecordReference === 0n