What causes "iflow process failed: %w" in chenhg5/cc-connect?

Inspect the wrapped waitErr (exit status/signal) with errors.Unwrap or %v of the cause

iflow process failed: %w

Error message

iflow process failed: %w

What it means

summarizeIFlowError returns this when it found no usable plain-text error line in the CLI output but the iflow process exited non-zero (waitErr != nil). The original exit error is wrapped as the cause.

Source

Thrown at agent/iflow/session.go:870

		for _, line := range strings.Split(stderrText, "\n") {
			line = strings.TrimSpace(line)
			if line == "" {
				continue
			}
			if strings.HasPrefix(line, "<Execution Info>") || strings.HasPrefix(line, "</Execution Info>") {
				continue
			}
			if strings.HasPrefix(line, "{") || strings.HasPrefix(line, "}") || strings.HasPrefix(line, "\"") {
				continue
			}
			if utf8.RuneCountInString(line) > 300 {
				line = string([]rune(line)[:300]) + "..."
			}
			return fmt.Errorf("%s", line)
		}
	}
	if waitErr != nil {
		return fmt.Errorf("iflow process failed: %w", waitErr)
	}
	return fmt.Errorf("iflow API request failed")
}

func (s *iflowSession) RespondPermission(_ string, _ core.PermissionResult) error {
	return nil
}

func (s *iflowSession) Events() <-chan core.Event {
	return s.events
}

func (s *iflowSession) CurrentSessionID() string {
	v, _ := s.sessionID.Load().(string)
	return v
}

func (s *iflowSession) Alive() bool {

View on GitHub (pinned to 4000b2338a)

Solutions

Inspect the wrapped waitErr (exit status/signal) with errors.Unwrap or %v of the cause
Run the same iflow command manually to reproduce and see full output
Check dmesg/logs for OOM or signal kills
Pin/rollback the iflow CLI version if a recent upgrade changed output format

Example fix

// before
err := session.Send(ctx, prompt, nil)
log.Print(err)
// after
if err := session.Send(ctx, prompt, nil); err != nil {
    log.Printf("iflow send failed: %v (cause: %v)", err, errors.Unwrap(err))
}

Defensive patterns

Strategy: try-catch

Try / catch

if err := session.Send(ctx, msg, nil); err != nil {
    if strings.Contains(err.Error(), "iflow process failed") {
        cause := errors.Unwrap(err)
        var ee *exec.ExitError
        if errors.As(cause, &ee) { log.Printf("exit code: %d", ee.ExitCode()) }
    }
}

Prevention

Pin the iflow CLI version to one whose output format is understood
Set memory/process limits to avoid OOM kills of the CLI
Capture full CLI output at debug level for post-mortem

When it happens

Trigger: readLoop calls summarizeIFlowError after the process died with a non-zero exit status while its output contained only JSON lines (or nothing parseable), e.g. a crash, signal, or silent failure.

Common situations: iflow CLI segfault or panic; OOM kill of the process; CLI updated to a new output format with no plain-text error line; process killed by timeout/exit signal.

Related errors

AI-assisted analysis of chenhg5/cc-connect@4000b2338a (2026-09-06). Data as JSON: /api/errors/2de5f0c4e566b049. Report an issue: GitHub.