multica-ai/multica · error

not authenticated

Error message

not authenticated

What it means

Thrown by autoWatchWorkspaces after `multica login` when the freshly saved profile config has an empty token. The function deliberately loads the exact profile via LoadCLIConfigForProfile (not the task-safe resolvers) and fails closed if authentication state is missing, aborting the post-login workspace auto-watch step.

Source

Thrown at server/cmd/multica/cmd_login.go:97

		return nil
	}

	fmt.Fprintf(os.Stderr, "\n→ Run 'multica daemon start' to start your local agent runtime.\n")
	return nil
}

func autoWatchWorkspaces(cmd *cobra.Command) error {
	// runLogin has already passed the human/local command guard and saved the
	// newly authenticated profile. Read that exact profile here rather than the
	// general task-safe resolvers, which intentionally fail closed on a lone
	// MULTICA_DAEMON_PORT signal.
	profile := resolveProfile(cmd)
	cfg, err := cli.LoadCLIConfigForProfile(profile)
	if err != nil {
		return err
	}
	if cfg.Token == "" {
		return fmt.Errorf("not authenticated")
	}
	if cfg.ServerURL == "" {
		return fmt.Errorf("server URL not configured")
	}

	client := cli.NewAPIClient(normalizeAPIBaseURL(cfg.ServerURL), "", cfg.Token)
	ctx, cancel := cli.APIContext(context.Background())
	defer cancel()

	var workspaces []struct {
		ID   string `json:"id"`
		Name string `json:"name"`
	}
	if err := client.GetJSON(ctx, "/api/workspaces", &workspaces); err != nil {
		return fmt.Errorf("list workspaces: %w", err)
	}

	if len(workspaces) == 0 {

View on GitHub (pinned to 2c0912b6ec)

Solutions

1. Inspect the profile config file for the active profile and confirm the token field is populated
2. Verify write permissions on the CLI config directory; fix the filesystem issue and re-run `multica login`
3. Ensure --profile matches the profile you authenticated against
4. Re-run `multica login` after correcting the environment

# bash: confirm the profile config actually holds a token after login
multica config show --output json 2>/dev/null | jq -e '.token | length > 0' >/dev/null \
  || { echo "login did not persist a token — check config dir permissions" >&2; exit 2; }

When it happens

Trigger: Login flow completes but the token fails to persist to the profile config file; the config file is unreadable/corrupted in a way that yields an empty Token; a profile flag pointing at a profile that was never actually written; filesystem permissions blocking the config write.

Common situations: Home or config directory read-only (containers, restricted CI); XDG/config path misconfiguration; concurrent logins overwriting the profile; auth backend returning success without a token.

Understand the failure class

• Authentication and authorization failures — expired tokens, bad credentials, and missing scopes.

Related errors

• server URL not configured
• get source agent: %w
• invalid token format: must start with %s
• unsupported platform: %s
• could not start the local login callback server (used to rec

AI-assisted analysis of multica-ai/multica@2c0912b6ec (2026-08-15). Data as JSON: /api/errors/7841c34d940e3462. Report an issue: GitHub.