paperclipai/paperclip · error · Error

workspace_durable_seed_missing

workspace_durable_seed_missing

Error message

workspace_durable_seed_missing

What it means

Guard in prepareSandboxManagedRuntime: the spec references a durable workspace seed archive, but the expected seed file is not present at runtime root before the workspace tarballs are built/extracted. The durable seed that should seed the workspace is missing entirely (as opposed to corrupted, which raises workspace_durable_seed_digest_mismatch).

Solutions

1. Regenerate the durable seed archive for the workspace and re-run the prepare
2. Fix the seed path in the runtime spec if it points at the wrong location
3. Fall back to a full workspace sync (non-seeded) if the seed is optional for the run

When it happens

Trigger: Thrown at packages/adapter-utils/src/sandbox-managed-runtime.ts:1096 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of paperclipai/paperclip@3f1d897a7c (2026-09-10). Data as JSON: /api/errors/e4a1d902a45a4ca9. Report an issue: GitHub.

Appendix: source

Thrown at packages/adapter-utils/src/sandbox-managed-runtime.ts:1110

  // access failure still fails the preparation: a workspace that exists but
  // cannot be read must not silently become an empty remote workspace.
  const syncWorkspace = input.syncWorkspace !== false &&
    (await fs.access(input.workspaceLocalDir).then(
      () => true,
      (error: NodeJS.ErrnoException) => {
        if (error.code === "ENOENT") return false;
        throw error;
      },
    ));
  const workspaceInboundMode = input.workspaceInboundMode ?? "host_current";
  const stageWorkspace =
    syncWorkspace && workspaceInboundMode !== "adopt_remote";
  const prepareWorkspaceSeed =
    syncWorkspace &&
    workspaceInboundMode === "adopt_remote" &&
    input.workspaceDurableSeed !== undefined;
  if (workspaceInboundMode === "durable_seed" && !input.workspaceDurableSeed) {
    throw new Error("workspace_durable_seed_missing");
  }

  // Reject any unsafe asset key before an archive path or an asset directory is
  // built from it. This runs before the git snapshot work so a bad key fails fast.
  for (const asset of input.assets ?? []) {
    assertRuntimeAssetKeyIsSafe(asset.key);
  }

  // Wrap a host-side staging sub-step or one scheduler task in its own span when
  // the caller injects a runtime span runner. The runner defaults to a no-op, so
  // a caller with no injected runner keeps the current control flow, and a
  // throwing runner never changes it (see `createRuntimeSpanRunner`). The two
  // pre-`pack` operations — the git enumeration and the baseline content-hash
  // walk — parent under the `stage.sync` step, so they stop showing up as a
  // hidden gap at the head of the step. Each inbound task (`stage.workspace`,
  // `stage.asset.<key>`, `stage.project.<id>`) and each outbound restore task
  // (`restore.workspace`, `restore.asset.<key>`) opens its own span, so two
  // concurrent tasks produce overlapping spans and the `pack` span nests under

View on GitHub (pinned to 3f1d897a7c)