Macro is a unified workspace for teams: email, chat, docs, tasks, agents, calls, and CRM — @-linked together with shared AI memory.
TypeScript agentaiai-agentsall-in-onecrmcrm-systememaillinearmcpmessagingnotesnotionofficerustslackslack-alternativestartupstartup-toolstypescriptworkspace
//! In-process agent-loop adapter for channel bot responses.
use std::sync::Arc;
use agent::{AgentLoop, StreamPart, to_rig_messages};
use ai_tools::{AiToolSet, ToolServiceContext, ToolSetWithPrompt};
use async_trait::async_trait;
use futures::StreamExt;
use macro_user_id::user_id::MacroUserIdStr;
use crate::domain::ports::AgentResponder;
/// [`AgentResponder`] backed by the in-process agent loop and AI toolset.
pub struct AgentLoopResponder {
agent_loop: AgentLoop,
tool_context: Arc<ToolServiceContext>,
toolset: Arc<AiToolSet>,
system_prompt: String,
}
impl AgentLoopResponder {
/// Create a responder from a pre-configured tool context and toolset.
pub fn new(tool_context: ToolServiceContext, tools: ToolSetWithPrompt) -> Self {
Self {
agent_loop: AgentLoop::new(tool_context.recorder.clone()),
tool_context: Arc::new(tool_context),
toolset: tools.toolset,
// The channel rules go last so they are the freshest instructions the
// model reads; ahead of the tool prompt they were often ignored.
system_prompt: format!("{}\n{}", tools.prompt, prompt::channel_mention::PROMPT),
}
}
}
#[async_trait]
impl AgentResponder for AgentLoopResponder {
#[tracing::instrument(skip(self, prompt, user_id), err)]
async fn respond(&self, user_id: &str, prompt: String) -> anyhow::Result<String> {
let user_id = MacroUserIdStr::try_from(user_id.to_string())?;
let toolset: Arc<dyn ai_toolset::ToolSet<ToolServiceContext> + Send + Sync> =
self.toolset.clone();
let usage_ctx = ai_usage::UsageContext::new(ai_usage::AiFeature::ChannelBot, user_id);
// Carry the feature on the context so tool-spawned subagents attribute to it.
let mut tool_context = (*self.tool_context).clone();
tool_context.usage_context = usage_ctx.clone();
let mut session = self
.agent_loop
.session(
toolset,
Arc::new(tool_context),
&self.system_prompt,
usage_ctx,
)
.await;
let messages = vec![agent::types::ChatMessage {
role: agent::types::Role::User,
content: agent::types::ChatMessageContent::Text(prompt),
attachments: None,
}];
let mut stream = session.send_message(to_rig_messages(&messages)).await?;
let mut text = String::new();
while let Some(part) = stream.next().await {
if let StreamPart::Content(chunk) = part? {
text.push_str(&chunk);
}
}
Ok(text.trim().to_string())
}
}