What causes "OpenCode executor error: {err}" in BloopAI/vibe-kanban?

Open the task log for the full message after 'OpenCode executor error:' to identify the failing API call or provider error.

BloopAI/vibe-kanban · error

OpenCode executor error: {err}

Error message

OpenCode executor error: {err}

What it means

After the OpenCode server starts successfully, spawn_inner runs the session (or slash command) via run_session/run_slash_command. Any Err from these — API failures, HTTP errors against the opencode server, cancelled/failed turns — is logged as "OpenCode executor error: {err}" and mapped to ExecutorExitResult::Failure sent over the exit signal, marking the task attempt as failed.

Source

Thrown at crates/executors/src/executors/opencode.rs:238

                auto_approve,
                server_password,
                models_cache_key,
                commit_reminder,
                commit_reminder_prompt,
                repo_context,
            };

            let result = match slash_command {
                Some(command) => {
                    run_slash_command(config, log_writer.clone(), command, cancel_for_task).await
                }
                None => run_session(config, log_writer.clone(), cancel_for_task).await,
            };
            let exit_result = match result {
                Ok(()) => ExecutorExitResult::Success,
                Err(err) => {
                    let _ = log_writer
                        .log_error(format!("OpenCode executor error: {err}"))
                        .await;
                    ExecutorExitResult::Failure
                }
            };
            let _ = exit_signal_tx.send(exit_result);
        });

        Ok(SpawnedChild {
            child,
            exit_signal: Some(exit_signal_rx),
            cancel: Some(cancel),
        })
    }

    /// Transform raw model data into ModelInfo structs.
    fn transform_models(
        &self,
        models: &std::collections::HashMap<String, ProviderModelInfo>,

View on GitHub (pinned to 4deb7eca8f)

Solutions

Open the task log for the full message after 'OpenCode executor error:' to identify the failing API call or provider error.
Verify provider credentials/auth for the configured model (`opencode auth` or environment API keys) and re-authenticate if expired.
Check the executor's configured model/variant/agent is valid for the installed opencode version; correct it in the task profile settings.
Retry the task if the failure was transient (rate limit, network blip); check provider status.
If the local server connection failed, ensure no firewall/proxy blocks localhost HTTP to the server's base_url.

Example fix

// before: executor profile with invalid model
OPENCODE_MODEL = "nonexistent-provider/gpt-x"
// after
OPENCODE_MODEL = "anthropic/claude-sonnet-4" // valid, authenticated provider/model

Defensive patterns

Strategy: try-catch

Validate before calling

// validate model/agent config before starting the task
if (!model.contains('/')) {
    return Err("invalid OPENCODE_MODEL, expected provider/model format".into());
}

Try / catch

match run_session(config, log_writer.clone(), cancel_for_task).await {
    Ok(()) => ExecutorExitResult::Success,
    Err(err) => {
      log_error(format!("OpenCode executor error: {err}"));
      // inspect err: auth failure -> re-auth; rate limit -> retry later; bad model -> fix profile
      ExecutorExitResult::Failure
    }
}

Prevention

Keep provider API keys/opencode auth current; re-authenticate before long-running tasks.
Validate configured model ids and variants against the provider list before launching.
Watch for rate limits and add spacing/retries for bulk task runs.
Confirm no local firewall/proxy blocks HTTP to the opencode server's base_url.
Read the executor log detail after 'OpenCode executor error:' to target the exact failing call.

When it happens

Trigger: run_session or run_slash_command (reached from spawn or spawn_follow_up) returns Err: the HTTP API calls to the local opencode server fail (connection refused, 4xx/5xx), the model/provider request errors, authentication fails, or the session is aborted with an error.

Common situations: Missing or expired provider API key / opencode auth; invalid model id or variant configured for the executor; opencode server rejecting requests (wrong credentials like OPENCODE_SERVER_PASSWORD mismatch); network/proxy issues calling provider APIs; prompt rejected by the model provider (rate limit, context length).

Related errors

AI-assisted analysis of BloopAI/vibe-kanban@4deb7eca8f (2026-08-29). Data as JSON: /api/errors/290f279d7fbd71fa. Report an issue: GitHub.