Hmbown/CodeWhale · error · anyhow::Error

unavailable credential

Error message

unavailable credential

What it means

Second site of the deliberate "unavailable credential" masking: while handling `auth print-api-key`, ConfigStore::load itself failed (invalid TOML, unreadable file, bad permissions) and the real error is replaced so config internals never leak through the credential pipe. The underlying problem is a broken config file, not missing credentials.

Source

Thrown at crates/cli/src/lib.rs:1761

        auth_mode: None,
        output_mode: cli.output_mode.clone(),
        log_level: cli.log_level.clone(),
        telemetry: cli.telemetry,
        approval_policy: cli.approval_policy.clone(),
        sandbox_mode: cli.sandbox_mode.clone(),
        yolo: Some(cli.yolo),
        verbosity: cli.verbosity.clone(),
    };
    if uses_raw_tui_provider
        && let Some((resolved_runtime, passthrough)) =
            prepare_raw_provider_tui_dispatch(&cli, command.as_ref(), &runtime_overrides)?
    {
        return run_tui_in_process(&cli, &resolved_runtime, passthrough);
    }

    let mut store = ConfigStore::load(cli.config.clone()).map_err(|error| {
        if pipe_api_key_handoff {
            anyhow!("unavailable credential")
        } else {
            error
        }
    })?;
    match command {
        Some(Commands::Run(args)) => {
            let resolved_runtime = resolve_runtime_for_dispatch(&mut store, &runtime_overrides);
            run_tui_in_process(&cli, &resolved_runtime, args.args)
        }
        Some(Commands::Doctor(args)) => {
            let resolved_runtime =
                resolve_runtime_for_diagnostic_dispatch(&store, &runtime_overrides);
            run_tui_in_process(&cli, &resolved_runtime, tui_args("doctor", args))
        }
        Some(Commands::Models(args)) => {
            let resolved_runtime = resolve_runtime_for_dispatch(&mut store, &runtime_overrides);
            run_tui_in_process(&cli, &resolved_runtime, tui_args("models", args))
        }

View on GitHub (pinned to 8880682c63)

Solutions

1. Run any non-handoff command (`codewhale doctor`) outside the pipe — it prints the true config-load error
2. Fix or restore the config file's TOML
3. As a workaround, pass --config pointing at a known-good file for the handoff

Example fix

# before
 codewhale auth print-api-key | my-client   # -> unavailable credential (config broken)

# after
 codewhale doctor                            # shows real error: invalid TOML at line N 
 # fix config, then: 
 codewhale auth print-api-key | my-client

# Validate the config outside the credential pipe before scripting the handoff:
 codewhale doctor >/dev/null 2>&1 || { echo "config broken — fix first"; exit 1; } 
 codewhale auth print-api-key | my-client

When it happens

Trigger: Running `codewhale auth print-api-key | client` while the config file fails to load: TOML syntax error from hand editing, file unreadable (permissions), or config written by an incompatible newer version.

Common situations: Hand-edited config with a missing quote/bracket; config shared across CLI versions; read-only $HOME in CI or containers.

Related errors

• unavailable credential
• Secret storage write failed for {slot}: {err}. Refusing to w
• {error}; additionally could not verify secret-store rollback
• {error}; additionally failed to restore prior secret-store s
• invalid JSON payload at key {MCP_SERVER_DEFINITIONS_KEY}; co

AI-assisted analysis of Hmbown/CodeWhale@8880682c63 (2026-08-16). Data as JSON: /api/errors/9d72903d8474a35f. Report an issue: GitHub.