coleam00/Archon · error · WorkflowRunFailedError

result.error

Error message

result.error

What it means

When a workflow completes unsuccessfully, the CLI throws `WorkflowRunFailedError(result.error, detachedProcessOwner)` (packages/cli/src/utils/workflow-exit-code.ts:8), carrying the workflow's error string and the detached flag; `resolveCliExitCode` maps it to a distinct exit code (including DETACHED_RUN_FAILED_EXIT_CODE) so wrappers can detect failure programmatically.

Source

Thrown at packages/cli/src/commands/workflow.ts:3213

    // Surface workflow result to Web UI as a result card (mirrors orchestrator.ts result message).
    // Paused workflows are handled in the branch above and intentionally do not get a result card.
    if ('summary' in result && result.summary) {
      try {
        await adapter.sendMessage(conversationId, result.summary, {
          category: 'workflow_result',
          segment: 'new',
          workflowResult: { workflowName: workflow.name, runId: result.workflowRunId },
        });
      } catch (surfaceError) {
        getLog().warn(
          { err: surfaceError as Error, conversationId },
          'cli.workflow_result_surface_failed'
        );
      }
    }
    console.log('\nWorkflow completed successfully.');
  } else {
    throw new WorkflowRunFailedError(result.error, detachedProcessOwner);
  }
}

/**
 * Run a specific workflow.
 *
 * A thin owner around the implementation: whatever the run does with its captured source,
 * the capture is either adopted by a run or reclaimed. The implementation has a dozen
 * ordinary ways out — unknown workflow, refused inputs, flag conflicts, a detached
 * dispatch — and asking each to remember a disposal call is how most of them did not.
 */
export async function workflowRunCommand(
  cwd: string,
  workflowName: string,
  userMessage: string,
  options: WorkflowRunOptions = {}
): Promise<void> {
  try {

View on GitHub (pinned to 0773b97458)

Solutions

1. Read the workflow's error text (from result.error, shown in output) and fix the underlying node-level cause.
2. Re-run the workflow after fixing; completed nodes are reused on resume where supported.
3. In detached mode, inspect the run-control status/log for the failed run before retrying.
4. Wrap invocations expecting possible failure: catch `WorkflowRunFailedError` and map its exit code via `resolveCliExitCode`.

Example fix

// before
try { await runWorkflowCommand(...) } catch (e) { console.error(e) }
// after
import { WorkflowRunFailedError, resolveCliExitCode } from '../utils/workflow-exit-code';
try {
  await runWorkflowCommand(...);
} catch (e) {
  if (e instanceof WorkflowRunFailedError) {
    console.error(`Workflow failed: ${e.message}`);
    process.exitCode = resolveCliExitCode(e);
  } else throw e;
}

function isWorkflowRunFailedError(e: unknown): e is WorkflowRunFailedError {
  return e instanceof WorkflowRunFailedError;
}

try {
  await runWorkflowCommand(...);
} catch (e) {
  if (isWorkflowRunFailedError(e)) {
    console.error(`Workflow failed: ${e.message}`);
    process.exitCode = resolveCliExitCode(e);
  } else {
    throw e;
  }
}

When it happens

Trigger: `executeWorkflow` returns `result.success === false`: a node failed, retries exhausted, a gate was rejected, or an infrastructure error failed the run — the CLI surfaces it as this typed error.

Common situations: A bash node exited non-zero; an AI node exhausted retries; a human gate was rejected; docker/network failure inside a node became a permanent run failure.

Related errors

• Cannot execute run '${detachedPreCreatedRun.id}': it belongs
• Dry-run failed; missing stubs: ${blockingMissingStubs.join('
• Failed to get workflow run: ${err.message}
• --open and --status are mutually exclusive: the inbox is fai
• Invalid --status '${opts.status}'. Valid: ${workflowRunStatu

AI-assisted analysis of coleam00/Archon@0773b97458 (2026-09-01). Data as JSON: /api/errors/bab68e765eeb42ad. Report an issue: GitHub.