@stigmer/runner 3.12.8 → 3.13.0

package/src/encryption/payload-codec.ts

@@ -1,152 +0,0 @@
-/**
- * Temporal PayloadCodec that encrypts payloads at rest in workflow
- * history (stigmer-cloud#227).
- *
- * Why: the workflow engine runs inside the Temporal deterministic
- * sandbox, so decrypted execution-context values cross the history
- * boundary in many places — the hydrate activity result, the runtime
- * env passed as an input to every per-task activity, and expression
- * results recorded as local-activity markers. Encrypting at the payload
- * codec layer closes the entire class with one mechanism instead of
- * chasing each crossing.
- *
- * Envelope (cross-SDK contract — the Java decode-only codec in
- * stigmer-cloud's temporal-starter must match it byte-for-byte, pinned
- * by the conformance fixture in __tests__/fixtures/):
- *
- *   metadata: encoding            = "binary/encrypted"
- *             encryption-key-id   = <key id that encrypted this payload>
- *   data:     iv (12 bytes) ‖ AES-256-GCM(ciphertext ‖ tag (16 bytes))
- *
- * The plaintext is the serialized ORIGINAL Payload proto (metadata AND
- * data), so decode restores the payload exactly — including its
- * original encoding — with no side channel.
- *
- * Decode passes through payloads it did not encode. This is what keeps
- * plaintext signals from the Java/Go orchestrators and pre-rollout
- * in-flight histories working with zero migration. Everything else
- * fails closed: unknown key id, missing key id, and ciphertext tampering
- * all throw rather than surfacing bogus payloads.
- */
-
-import { createCipheriv, createDecipheriv, randomBytes } from "node:crypto";
-import type { Payload, PayloadCodec } from "@temporalio/common";
-// Default-import + destructure, NOT `import { temporal } from …`:
-// @temporalio/proto is CommonJS and its `temporal` export defeats Node's
-// cjs-module-lexer named-export detection, so the named form loads under
-// tsx/vitest (their interop is looser) but crashes plain `node dist/main.js`
-// at startup with "Named export 'temporal' not found" (stigmer/stigmer#399
-// boot regression). Pinned by scripts/verify-dist-boot.mjs in CI.
-import proto from "@temporalio/proto";
-import type { EncryptionKey, PayloadEncryptionConfig } from "./config.js";
-
-const { temporal } = proto;
-
-const ENCODING_METADATA_KEY = "encoding";
-const ENCRYPTED_ENCODING_VALUE = "binary/encrypted";
-const KEY_ID_METADATA_KEY = "encryption-key-id";
-
-/** AES-GCM parameters shared with the Java implementation. */
-const IV_BYTES = 12;
-const AUTH_TAG_BYTES = 16;
-
-export class EncryptionPayloadCodec implements PayloadCodec {
-  private readonly decryptKeysById: Map<string, Buffer>;
-
-  constructor(private readonly config: PayloadEncryptionConfig) {
-    this.decryptKeysById = new Map([[config.primary.keyId, config.primary.key]]);
-    if (config.secondary) {
-      this.decryptKeysById.set(config.secondary.keyId, config.secondary.key);
-    }
-  }
-
-  async encode(payloads: Payload[]): Promise<Payload[]> {
-    return payloads.map((p) => this.encodePayload(p));
-  }
-
-  async decode(payloads: Payload[]): Promise<Payload[]> {
-    return payloads.map((p) => this.decodePayload(p));
-  }
-
-  private encodePayload(payload: Payload): Payload {
-    // Data-less payloads (binary/null from void results) stay as-is:
-    // there is nothing to protect, and the cross-language parents that
-    // await our workflows as void must be able to read them without a key.
-    if (!payload.data || payload.data.length === 0) {
-      return payload;
-    }
-
-    const plaintext = temporal.api.common.v1.Payload.encode(payload).finish();
-    return {
-      metadata: {
-        [ENCODING_METADATA_KEY]: Buffer.from(ENCRYPTED_ENCODING_VALUE),
-        [KEY_ID_METADATA_KEY]: Buffer.from(this.config.primary.keyId),
-      },
-      data: encrypt(plaintext, this.config.primary),
-    };
-  }
-
-  private decodePayload(payload: Payload): Payload {
-    if (!isEncryptedPayload(payload)) {
-      return payload;
-    }
-
-    const keyIdBytes = payload.metadata?.[KEY_ID_METADATA_KEY];
-    if (!keyIdBytes) {
-      throw new Error(
-        "Encrypted payload is missing its encryption-key-id metadata — refusing to decode",
-      );
-    }
-    const keyId = Buffer.from(keyIdBytes).toString("utf-8");
-    const key = this.decryptKeysById.get(keyId);
-    if (!key) {
-      throw new Error(
-        `Encrypted payload uses unknown key id '${keyId}' — configure it as the ` +
-          `primary or secondary payload encryption key (rotation window?)`,
-      );
-    }
-
-    const plaintext = decrypt(payload.data!, key, keyId);
-    return temporal.api.common.v1.Payload.decode(plaintext);
-  }
-}
-
-function isEncryptedPayload(payload: Payload): boolean {
-  const encoding = payload.metadata?.[ENCODING_METADATA_KEY];
-  if (!encoding) return false;
-  return Buffer.from(encoding).toString("utf-8") === ENCRYPTED_ENCODING_VALUE;
-}
-
-function encrypt(plaintext: Uint8Array, key: EncryptionKey): Buffer {
-  const iv = randomBytes(IV_BYTES);
-  const cipher = createCipheriv("aes-256-gcm", key.key, iv);
-  const ciphertext = Buffer.concat([cipher.update(plaintext), cipher.final()]);
-  // Layout iv ‖ ciphertext ‖ tag matches Java's AES/GCM/NoPadding, whose
-  // doFinal() output is ciphertext ‖ tag.
-  return Buffer.concat([iv, ciphertext, cipher.getAuthTag()]);
-}
-
-function decrypt(data: Uint8Array, key: Buffer, keyId: string): Buffer {
-  if (data.length < IV_BYTES + AUTH_TAG_BYTES) {
-    throw new Error(
-      `Encrypted payload under key id '${keyId}' is truncated (${data.length} bytes)`,
-    );
-  }
-  const buf = Buffer.from(data);
-  const iv = buf.subarray(0, IV_BYTES);
-  const ciphertext = buf.subarray(IV_BYTES, buf.length - AUTH_TAG_BYTES);
-  const tag = buf.subarray(buf.length - AUTH_TAG_BYTES);
-
-  const decipher = createDecipheriv("aes-256-gcm", key, iv);
-  decipher.setAuthTag(tag);
-  try {
-    return Buffer.concat([decipher.update(ciphertext), decipher.final()]);
-  } catch {
-    // GCM auth failure: tampered ciphertext or a key that does not match
-    // its advertised id. Never surface partially decrypted bytes.
-    throw new Error(
-      `Failed to decrypt payload under key id '${keyId}' — ciphertext is ` +
-        `corrupt or the configured key does not match`,
-    );
-  }
-}

package/src/shared/zip-structure.ts

@@ -1,181 +0,0 @@
-/**
- * Policy-free structural ZIP parsing: end-of-central-directory (EOCD)
- * location plus central-directory walk, producing one record per entry.
- *
- * Parsing is *central-directory-based* — never a front-to-back walk of
- * local file headers. This is a correctness decision, not a style choice
- * (issues #450 and #567): local headers of streaming entries
- * (general-purpose flag bit 3) carry zeroed sizes, and the only
- * local-only way to recover them is scanning the payload for the
- * data-descriptor signature — which silently truncates any stored entry
- * whose *content* happens to contain those four bytes, and
- * desynchronizes every entry after it. The central directory always
- * carries the real sizes, and every consumer here holds the complete
- * archive bytes, so nothing a local-header walk could offer is needed.
- *
- * This layer maps bytes to entry records and nothing else. Policy — what
- * a structural failure means, which entries are acceptable, how payloads
- * are decoded — belongs to the consumers, and they differ on purpose:
- *
- *   - shared/zip-extract.ts (skill artifacts): structural failure is
- *     NON-FATAL — both editions' push gates validated every artifact
- *     with a central-directory-based reader before storage, so a defect
- *     here can only be a truncated or corrupted download.
- *   - execute-deep-agent/attachment-injector.ts (user attachments):
- *     structural failure is FAIL-HARD — the input is an untrusted
- *     upload and nothing upstream vouched for it.
- *
- * ZIP64 is deliberately unsupported: both consumers cap input far below
- * every ZIP64 threshold (skill push gates: 100MB / 10,000 files;
- * attachment uploads: 10MB).
- *
- * Throws plain `Error`s on structural defects (no valid EOCD, a central
- * directory or local header record that does not parse, a payload slice
- * that runs past the end of the buffer); consumers translate those into
- * their own error models.
- */
-
-// ─── Entry records ───────────────────────────────────────────────────────
-
-export interface ZipStructuralEntry {
-  /** Entry path exactly as recorded in the central directory. */
-  readonly name: string;
-  readonly isDirectory: boolean;
-  readonly compressionMethod: number;
-  /**
-   * The central directory's declared uncompressed size — authoritative for
-   * pre-extraction accounting (e.g. ZIP-bomb budgeting), but still a
-   * *declaration*: a consumer that distrusts its input must enforce it
-   * against the actual decompressed output.
-   */
-  readonly uncompressedSize: number;
-  /** The entry's raw payload slice (a view, not a copy) of the archive buffer. */
-  readonly compressedData: Uint8Array;
-}
-
-// ─── Format constants ────────────────────────────────────────────────────
-
-const LOCAL_FILE_HEADER_SIG = 0x04034b50;
-const CENTRAL_DIRECTORY_SIG = 0x02014b50;
-const EOCD_SIG = 0x06054b50;
-
-/** Fixed size of the EOCD record, excluding the variable-length comment. */
-export const EOCD_MIN_SIZE = 22;
-/** The archive comment length is a u16, so the EOCD sits within the last 64KB + 22 bytes. */
-const EOCD_MAX_COMMENT = 0xffff;
-
-const LOCAL_HEADER_SIZE = 30;
-const CD_RECORD_SIZE = 46;
-
-// ─── Parsing ─────────────────────────────────────────────────────────────
-
-/**
- * Enumerate the archive's entries from its central directory.
- *
- * Throws on any structural defect: no valid EOCD, a central directory or
- * local header record that doesn't parse, or a payload slice that runs
- * past the end of the buffer.
- */
-export function parseZipStructure(data: Uint8Array): ZipStructuralEntry[] {
-  const view = new DataView(data.buffer, data.byteOffset, data.byteLength);
-  const eocd = findEndOfCentralDirectory(data, view);
-
-  const entries: ZipStructuralEntry[] = [];
-  let offset = eocd.centralDirectoryOffset;
-
-  for (let i = 0; i < eocd.entryCount; i++) {
-    if (offset + CD_RECORD_SIZE > data.length || view.getUint32(offset, true) !== CENTRAL_DIRECTORY_SIG) {
-      throw new Error(`invalid central directory record at offset ${offset}`);
-    }
-
-    const compressionMethod = view.getUint16(offset + 10, true);
-    const compressedSize = view.getUint32(offset + 20, true);
-    const uncompressedSize = view.getUint32(offset + 24, true);
-    const fileNameLength = view.getUint16(offset + 28, true);
-    const extraFieldLength = view.getUint16(offset + 30, true);
-    const commentLength = view.getUint16(offset + 32, true);
-    const localHeaderOffset = view.getUint32(offset + 42, true);
-
-    const fileName = new TextDecoder().decode(
-      data.subarray(offset + CD_RECORD_SIZE, offset + CD_RECORD_SIZE + fileNameLength),
-    );
-
-    // The local header is consulted for one thing only: the payload start.
-    // Its name/extra field lengths can legitimately differ from the central
-    // directory's (streaming writers pad the extra field), so the offset
-    // cannot be derived from CD fields alone.
-    const dataStart = payloadStart(data, view, localHeaderOffset, fileName);
-    if (dataStart + compressedSize > data.length) {
-      throw new Error(`entry "${fileName}" payload runs past end of archive`);
-    }
-
-    entries.push({
-      name: fileName,
-      isDirectory: fileName.endsWith("/"),
-      compressionMethod,
-      uncompressedSize,
-      compressedData: data.subarray(dataStart, dataStart + compressedSize),
-    });
-
-    offset += CD_RECORD_SIZE + fileNameLength + extraFieldLength + commentLength;
-  }
-
-  return entries;
-}
-
-interface EndOfCentralDirectory {
-  centralDirectoryOffset: number;
-  entryCount: number;
-}
-
-/**
- * Locate and validate the end-of-central-directory record.
- *
- * Scans backward from the end of the archive across the maximum comment
- * span. A signature match alone is not trusted (the four bytes can occur
- * inside a trailing comment): the record must also point at an offset
- * that actually holds a central directory record, or be a genuinely
- * empty archive.
- */
-function findEndOfCentralDirectory(data: Uint8Array, view: DataView): EndOfCentralDirectory {
-  const scanFloor = Math.max(0, data.length - EOCD_MIN_SIZE - EOCD_MAX_COMMENT);
-
-  for (let pos = data.length - EOCD_MIN_SIZE; pos >= scanFloor; pos--) {
-    if (view.getUint32(pos, true) !== EOCD_SIG) continue;
-
-    const entryCount = view.getUint16(pos + 10, true);
-    const centralDirectorySize = view.getUint32(pos + 12, true);
-    const centralDirectoryOffset = view.getUint32(pos + 16, true);
-
-    const directoryEndsAtRecord = centralDirectoryOffset + centralDirectorySize <= pos;
-    const directoryLooksReal =
-      entryCount === 0 ||
-      (centralDirectoryOffset + 4 <= data.length &&
-        view.getUint32(centralDirectoryOffset, true) === CENTRAL_DIRECTORY_SIG);
-
-    if (directoryEndsAtRecord && directoryLooksReal) {
-      return { centralDirectoryOffset, entryCount };
-    }
-  }
-
-  throw new Error("no end-of-central-directory record found");
-}
-
-/** Resolve where an entry's payload begins, from its local file header. */
-function payloadStart(
-  data: Uint8Array,
-  view: DataView,
-  localHeaderOffset: number,
-  fileName: string,
-): number {
-  if (
-    localHeaderOffset + LOCAL_HEADER_SIZE > data.length ||
-    view.getUint32(localHeaderOffset, true) !== LOCAL_FILE_HEADER_SIG
-  ) {
-    throw new Error(`entry "${fileName}" has no local file header at offset ${localHeaderOffset}`);
-  }
-
-  const localNameLength = view.getUint16(localHeaderOffset + 26, true);
-  const localExtraLength = view.getUint16(localHeaderOffset + 28, true);
-  return localHeaderOffset + LOCAL_HEADER_SIZE + localNameLength + localExtraLength;
-}