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/apkSignature/apkV1SignatureBlock.test.ts

@@ -202,9 +202,8 @@ describe("v1 signature block (PKCS#7 SignedData)", () => {
     let nested: Uint8Array = Uint8Array.from([0x05, 0x00]);
     for (let i = 0; i < 64; i += 1) nested = indefinite(0x30, nested);
     expect(errorOf(nested)).toEqual({
-      kind: "budget-exceeded",
-      resource: "asn1-depth",
-      limit: 32,
+      kind: "malformed-pkcs7",
+      detail: "ASN.1 nesting exceeds 32 levels.",
     });
   });
 });
@@ -412,42 +411,3 @@ it.each([0, 1])(
     });
   },
 );
-
-it("keeps a valid constructed CMS content depth refusal distinct from malformed input", () => {
-  let content = tlv(0x04, Uint8Array.of(1));
-  const block = () =>
-    signedDataBlock({
-      certificates: [LEAF_CERTIFICATE],
-      signerInfos: [signerInfo({ names: LEAF_CERTIFICATE })],
-      encapsulatedContent: content,
-    });
-  expect(decodeV1SignatureBlock(block()).ok).toBe(true);
-  for (let depth = 0; depth < 40; depth += 1)
-    content = indefinite(0x24, content);
-  expect(errorOf(block())).toEqual({
-    kind: "budget-exceeded",
-    resource: "asn1-depth",
-    limit: 32,
-  });
-});
-
-it("preserves the canonical OID byte allowance for a valid unknown digest algorithm", () => {
-  const algorithm = tlv(
-    0x06,
-    Uint8Array.from([0x2a, ...Array<number>(64).fill(0)]),
-  );
-  expect(
-    errorOf(
-      signedDataBlock({
-        certificates: [LEAF_CERTIFICATE],
-        signerInfos: [
-          signerInfo({
-            names: LEAF_CERTIFICATE,
-            digestAlgorithmOid: algorithm,
-          }),
-        ],
-        digestAlgorithmOid: algorithm,
-      }),
-    ),
-  ).toEqual({ kind: "budget-exceeded", resource: "oid-bytes", limit: 64 });
-});

src/domain/apkSignature/apkV1SignatureBlock.ts

@@ -27,10 +27,7 @@ export type ApkV1BlockError =
       readonly resource:
         | "v1-block-certificates"
         | "v1-signer-infos"
-        | "asn1-elements"
-        | "asn1-depth"
-        | "oid-bytes"
-        | "oid-arc-bytes";
+        | "asn1-elements";
       readonly limit: number;
     };
 
@@ -101,14 +98,11 @@ class ApkBerReader {
       return this.#reader.read(at, limit, depth);
     } catch (cause: unknown) {
       if (!(cause instanceof Asn1BerFailure)) throw cause;
-      if (cause.constraint !== null)
+      if (cause.kind === "budget-exceeded")
         throw new BlockFailure({
           kind: "budget-exceeded",
-          resource:
-            cause.constraint.resource === "nodes"
-              ? "asn1-elements"
-              : "asn1-depth",
-          limit: cause.constraint.limit,
+          resource: "asn1-elements",
+          limit: MAX_ASN1_ELEMENTS,
         });
       throw malformed(cause.message);
     }
@@ -161,13 +155,6 @@ const expectOctetString = (
 const oidText = (content: Uint8Array): string => {
   const decoded = decodeAsn1Oid(content, { bytes: 64, arcBytes: 64 });
   if (decoded.ok) return decoded.value;
-  if (decoded.error.kind === "byte-limit" || decoded.error.kind === "arc-limit")
-    throw new BlockFailure({
-      kind: "budget-exceeded",
-      resource:
-        decoded.error.kind === "byte-limit" ? "oid-bytes" : "oid-arc-bytes",
-      limit: decoded.error.limit,
-    });
   if (decoded.error.kind === "truncated")
     throw malformed("object identifier is truncated");
   throw malformed(`object identifier of ${String(content.length)} bytes`);
@@ -201,13 +188,6 @@ const integerIdentityContents = (bytes: Uint8Array): Uint8Array => {
   return start === 0 ? bytes : bytes.subarray(start);
 };
 
-/** Optional identity interpretation may remain unknown while raw certificate identity survives. */
-const unavailableIdentityInterpretation = (cause: unknown): boolean =>
-  cause instanceof BlockFailure &&
-  (cause.error.kind === "malformed-pkcs7" ||
-    cause.error.resource === "oid-bytes" ||
-    cause.error.resource === "oid-arc-bytes");
-
 /**
  * RFC 5280 §7.1: RDN order matters; attributes in one SET do not. This
  * comparison profile admits ASCII PrintableString/UTF8String for common
@@ -257,7 +237,8 @@ const directoryNameKey = (name: Asn1BerNode): string | null => {
     }
     return JSON.stringify(rdns);
   } catch (cause) {
-    if (unavailableIdentityInterpretation(cause)) return null;
+    if (cause instanceof BlockFailure && cause.error.kind === "malformed-pkcs7")
+      return null;
     throw cause;
   }
 };
@@ -317,7 +298,8 @@ const certificateIdentity = (
       subjectKeyIdentifier,
     };
   } catch (cause: unknown) {
-    if (unavailableIdentityInterpretation(cause)) return null;
+    if (cause instanceof BlockFailure && cause.error.kind === "malformed-pkcs7")
+      return null;
     throw cause;
   }
 };

src/domain/asn1Ber.ts

@@ -11,16 +11,11 @@ export interface Asn1BerNode {
   readonly children: readonly Asn1BerNode[];
 }
 
-/** The fixed parser allowance that prevented examining more BER structure. */
-export interface Asn1BerConstraint {
-  readonly resource: "nodes" | "depth";
-  readonly limit: number;
-}
-
 export class Asn1BerFailure extends Error {
   constructor(
     message: string,
-    readonly constraint: Asn1BerConstraint | null = null,
+    readonly kind: "malformed" | "budget-exceeded" = "malformed",
+    readonly limit: number | null = null,
   ) {
     super(message);
     this.name = "Asn1BerFailure";
@@ -33,9 +28,10 @@ interface Budget {
 
 const fail = (
   message: string,
-  constraint: Asn1BerConstraint | null = null,
+  kind: Asn1BerFailure["kind"] = "malformed",
+  limit: number | null = null,
 ): never => {
-  throw new Asn1BerFailure(message, constraint);
+  throw new Asn1BerFailure(message, kind, limit);
 };
 
 /**
@@ -55,14 +51,12 @@ export class Asn1BerReader {
 
   read(offset = 0, end = this.bytes.length, depth = 0): Asn1BerNode {
     if (depth > this.maxDepth)
-      return fail(`ASN.1 nesting exceeds ${String(this.maxDepth)} levels.`, {
-        resource: "depth",
-        limit: this.maxDepth,
-      });
+      return fail(`ASN.1 nesting exceeds ${String(this.maxDepth)} levels.`);
     if (this.#budget.remaining <= 0)
       return fail(
         `ASN.1 value count exceeds the ${String(this.maxNodes)}-node budget.`,
-        { resource: "nodes", limit: this.maxNodes },
+        "budget-exceeded",
+        this.maxNodes,
       );
     this.#budget.remaining -= 1;
     if (offset >= end || offset + 2 > end)

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/native/pe/PeCertificateAsn1.ts

@@ -28,12 +28,7 @@ const OID_NAMES: Readonly<Record<string, string>> = {
 
 /** A fixed per-payload decoder allowance and its limit. */
 export interface PeCertificateDecodeAllowance {
-  readonly allowance:
-    | "asn1-nodes"
-    | "asn1-depth"
-    | "decoded-text"
-    | "oid-bytes"
-    | "oid-arc-bytes";
+  readonly allowance: "asn1-nodes" | "decoded-text";
   readonly limit: number;
 }
 
@@ -67,15 +62,9 @@ export const readAsn1 = (bytes: Buffer): Asn1BerNode => {
     if (cause instanceof Asn1BerFailure)
       throw new PeCertificateAsn1Failure(
         cause.message,
-        cause.constraint === null
-          ? null
-          : {
-              allowance:
-                cause.constraint.resource === "nodes"
-                  ? "asn1-nodes"
-                  : "asn1-depth",
-              limit: cause.constraint.limit,
-            },
+        cause.kind === "budget-exceeded"
+          ? { allowance: "asn1-nodes", limit: cause.limit ?? ASN1_NODE_BUDGET }
+          : null,
       );
     throw cause;
   }
@@ -123,14 +112,12 @@ export const decodeOid = (content: Buffer): string => {
     case "empty":
       return fail("Empty OBJECT IDENTIFIER.");
     case "byte-limit":
-      throw new PeCertificateAsn1Failure(
+      return fail(
         `OBJECT IDENTIFIER exceeds ${String(decoded.error.limit)} bytes.`,
-        { allowance: "oid-bytes", limit: decoded.error.limit },
       );
     case "arc-limit":
-      throw new PeCertificateAsn1Failure(
+      return fail(
         `OBJECT IDENTIFIER arc exceeds ${String(decoded.error.limit)} bytes.`,
-        { allowance: "oid-arc-bytes", limit: decoded.error.limit },
       );
     case "non-minimal":
       return fail("Non-minimal OBJECT IDENTIFIER component.");

src/native/pe/PeCertificatesProvider.test.ts

@@ -1,132 +0,0 @@
-import { writeFile } from "node:fs/promises";
-import { join } from "node:path";
-import { expect, it } from "vitest";
-import { createTestTempDirectory } from "../../../tests/fixtures/temporaryDirectory.js";
-import { repeatedCertificateSignedData } from "../../../tests/fixtures/peCertificateAllowance.js";
-import {
-  indefinite,
-  tlv,
-  signedDataBlock,
-  signerInfo,
-  LEAF_CERTIFICATE,
-} from "../../../tests/fixtures/apkSignature/v1SignatureBlock.js";
-import { inspectPeCertificatesInputSchema } from "../../domain/native/peCertificates.js";
-import {
-  AnalysisInputError,
-  AnalysisResourceConstraintError,
-} from "../../domain/analysisErrorCore.js";
-import { peCertificateFixture } from "./PeCertificates.fixture.js";
-import { PeCertificatesProvider } from "./PeCertificatesProvider.js";
-
-const encapsulated = (depth: number): Buffer => {
-  let content = tlv(0x04, Uint8Array.of(1));
-  for (let level = 0; level < depth; level += 1)
-    content = indefinite(0x24, content);
-  return Buffer.from(
-    signedDataBlock({
-      certificates: [LEAF_CERTIFICATE],
-      signerInfos: [signerInfo({ names: LEAF_CERTIFICATE })],
-      encapsulatedContent: content,
-    }),
-  );
-};
-
-const inspect = async (payload: Buffer) => {
-  const path = join(
-    await createTestTempDirectory("rea-pe-decoder-allowance-"),
-    "inert.exe",
-  );
-  await writeFile(path, peCertificateFixture(false, payload));
-  const provider = new PeCertificatesProvider();
-  try {
-    return await provider.inspect(
-      inspectPeCertificatesInputSchema.parse({ path }),
-    );
-  } finally {
-    await provider.close();
-  }
-};
-
-it.each([
-  {
-    name: "depth",
-    payload: () => encapsulated(40),
-    limits: { max_asn1_depth: 32 },
-    description: "32 nesting levels",
-  },
-  {
-    name: "nodes",
-    payload: () => repeatedCertificateSignedData(600),
-    limits: { max_asn1_values: 16_384 },
-    description: "16384 values",
-  },
-  {
-    name: "text",
-    payload: () => repeatedCertificateSignedData(350),
-    limits: { max_decoded_text_bytes: 65_536 },
-    description: "65536 bytes",
-  },
-])(
-  "reports the fixed $name allowance with its own resource and recovery",
-  async ({ payload, limits, description }) => {
-    const result = await inspect(payload());
-    expect(result.ok).toBe(false);
-    if (result.ok) throw new Error("Expected decoder refusal");
-    expect(result.error).toBeInstanceOf(AnalysisResourceConstraintError);
-    if (!(result.error instanceof AnalysisResourceConstraintError))
-      throw result.error;
-    expect(result.error.reportedLimits).toEqual(limits);
-    expect(result.error.retryable).toBe(false);
-    expect(result.error.fixedAllowance?.name).toContain(description);
-    expect(result.error.fixedAllowance?.name).toContain("no runtime override");
-  },
-);
-
-it("accepts shallower valid BER and keeps malformed BER an input failure", async () => {
-  expect((await inspect(encapsulated(1))).ok).toBe(true);
-  const malformed = await inspect(Buffer.from([0x30, 0x03, 0x05]));
-  expect(malformed.ok).toBe(false);
-  if (malformed.ok) throw new Error("Expected malformed input refusal");
-  expect(malformed.error).toBeInstanceOf(AnalysisInputError);
-});
-
-it.each([
-  {
-    name: "bytes",
-    content: Buffer.from([0x2a, ...Array<number>(256).fill(0)]),
-    limits: { max_oid_bytes: 256 },
-  },
-  {
-    name: "arc bytes",
-    content: Buffer.from([0x2a, 0x81, ...Array<number>(31).fill(0x80), 0]),
-    limits: { max_oid_arc_bytes: 32 },
-  },
-])(
-  "preserves the canonical OID $name allowance in a valid CMS algorithm",
-  async ({ content, limits }) => {
-    const algorithm = tlv(0x06, content);
-    const result = await inspect(
-      Buffer.from(
-        signedDataBlock({
-          certificates: [LEAF_CERTIFICATE],
-          signerInfos: [
-            signerInfo({
-              names: LEAF_CERTIFICATE,
-              digestAlgorithmOid: algorithm,
-            }),
-          ],
-          digestAlgorithmOid: algorithm,
-        }),
-      ),
-    );
-    expect(result.ok).toBe(false);
-    if (result.ok) throw new Error("Expected OID allowance refusal");
-    expect(result.error).toBeInstanceOf(AnalysisResourceConstraintError);
-    if (!(result.error instanceof AnalysisResourceConstraintError))
-      throw result.error;
-    expect(result.error.reportedLimits).toEqual(limits);
-    expect(result.error.retryable).toBe(false);
-    expect(result.error.fixedAllowance?.name).toContain("OBJECT IDENTIFIER");
-    expect(result.error.fixedAllowance?.name).toContain("no runtime override");
-  },
-);

src/native/pe/PeCertificatesProvider.ts

@@ -105,40 +105,24 @@ const inputBudgetRemediation = (
 const decoderAllowanceFailure = (
   cause: PeCertificateAsn1Failure,
   { allowance, limit }: PeCertificateDecodeAllowance,
-) => {
-  const selected = {
-    "asn1-nodes": {
-      limits: { max_asn1_values: limit },
-      description: `The PE certificate ASN.1 allowance of ${String(limit)} values per PKCS#7 payload (this release has no runtime override)`,
-    },
-    "asn1-depth": {
-      limits: { max_asn1_depth: limit },
-      description: `The PE certificate ASN.1 allowance of ${String(limit)} nesting levels per PKCS#7 payload (this release has no runtime override)`,
-    },
-    "decoded-text": {
-      limits: { max_decoded_text_bytes: limit },
-      description: `The PE certificate decoded-text allowance of ${String(limit)} bytes per PKCS#7 payload (this release has no runtime override)`,
-    },
-    "oid-bytes": {
-      limits: { max_oid_bytes: limit },
-      description: `The PE certificate OBJECT IDENTIFIER allowance of ${String(limit)} bytes per OID (this release has no runtime override)`,
-    },
-    "oid-arc-bytes": {
-      limits: { max_oid_arc_bytes: limit },
-      description: `The PE certificate OBJECT IDENTIFIER allowance of ${String(limit)} bytes per base-128 arc (this release has no runtime override)`,
-    },
-  }[allowance];
-  return new AnalysisResourceConstraintError(
+) =>
+  new AnalysisResourceConstraintError(
     OPERATION,
     "memory",
     `PKCS#7 signedData exceeds a fixed decoder allowance: ${cause.message}`,
-    selected.limits,
+    allowance === "asn1-nodes"
+      ? { max_asn1_values: limit }
+      : { max_decoded_text_bytes: limit },
     {
       cause,
-      ...fixedAllowanceOptions(selected.description, "parser"),
+      ...fixedAllowanceOptions(
+        allowance === "asn1-nodes"
+          ? `The PE certificate ASN.1 allowance of ${String(limit)} values per PKCS#7 payload (this release has no runtime override)`
+          : `The PE certificate decoded-text allowance of ${String(limit)} bytes per PKCS#7 payload (this release has no runtime override)`,
+        "parser",
+      ),
     },
   );
-};
 
 const peCertificatesFailurePrimary = (
   facts: ArtifactFilePrimaryFacts,

tests/fixtures/apkSignature/v1SignatureBlock.ts

@@ -230,15 +230,12 @@ export interface SignerInfoSpec {
    * from the certificate named by `names`.
    */
   readonly serial?: Uint8Array;
-  /** Complete digest algorithm OID TLV, including bounded-parser fixtures. */
-  readonly digestAlgorithmOid?: Uint8Array;
 }
 
 export const signerInfo = ({
   names,
   signature,
   serial,
-  digestAlgorithmOid,
 }: SignerInfoSpec): Uint8Array =>
   tlv(
     0x30,
@@ -247,7 +244,7 @@ export const signerInfo = ({
       names,
       serial === undefined ? undefined : tlv(0x02, serial),
     ),
-    tlv(0x30, digestAlgorithmOid ?? oid(OID_SHA256)),
+    algorithm(OID_SHA256),
     algorithm(OID_ECDSA_SHA256),
     signature ?? tlv(0x04, Uint8Array.from([1, 2, 3])),
   );
@@ -257,8 +254,6 @@ export const signedDataBlock = (options: {
   readonly certificates: readonly Uint8Array[];
   readonly signerInfos: readonly Uint8Array[];
   readonly contentType?: string;
-  readonly encapsulatedContent?: Uint8Array;
-  readonly digestAlgorithmOid?: Uint8Array;
 }): Uint8Array =>
   tlv(
     0x30,
@@ -268,14 +263,8 @@ export const signedDataBlock = (options: {
       tlv(
         0x30,
         integer(1),
-        tlv(0x31, tlv(0x30, options.digestAlgorithmOid ?? oid(OID_SHA256))),
-        tlv(
-          0x30,
-          oid("1.2.840.113549.1.7.1"),
-          ...(options.encapsulatedContent === undefined
-            ? []
-            : [tlv(0xa0, options.encapsulatedContent)]),
-        ),
+        tlv(0x31, algorithm(OID_SHA256)),
+        tlv(0x30, oid("1.2.840.113549.1.7.1")),
         tlv(0xa0, ...options.certificates),
         tlv(0x31, ...options.signerInfos),
       ),