wangjunxiao/nanoclaw
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(permissions): welcome new approved channels via /welcome, route to them

Browse Source 
When the unknown-channel approval flow completes, seed a /welcome task
into the newly-wired session so the agent greets the new user on first
contact. The replayed /start (Telegram's default first-message) is filtered
by the agent-runner's command-command filter, so without an explicit
onboarding trigger the first interaction produced nothing.

Pin the destination by its local_name from agent_destinations to avoid the
agent picking the wrong named destination (previously it greeted the owner,
whose DM is in CLAUDE.md).

Also guard dispatchResultText against echoing trailing status lines when
the agent has already sent messages explicitly via send_message. Otherwise
a task-triggered flow that calls send_message then emits "welcome message
sent" produces a duplicate chat to the recipient.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Koshkoshinsk
2026-04-21 11:40:12 +00:00
parent 483969a194
commit 9776dd4f32
3 changed files with 181 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

35
container/agent-runner/src/poll-loop.ts
View File

@@ -1,7 +1,7 @@
import { findByName, getAllDestinations, type DestinationEntry } from './destinations.js';
import { getPendingMessages, markProcessing, markCompleted, type MessageInRow } from './db/messages-in.js';
import { writeMessageOut } from './db/messages-out.js';
import { touchHeartbeat, clearStaleProcessingAcks } from './db/connection.js';
import { touchHeartbeat, clearStaleProcessingAcks, getOutboundDb } from './db/connection.js';
import { getStoredSessionId, setStoredSessionId, clearStoredSessionId } from './db/session-state.js';
import { formatMessages, extractRouting, categorizeMessage, stripInternalTags, type RoutingContext } from './formatter.js';
import type { AgentProvider, AgentQuery, ProviderEvent } from './providers/types.js';
@@ -280,6 +280,17 @@ async function processQuery(query: AgentQuery, routing: RoutingContext): Promise
let queryContinuation: string | undefined;
let done = false;
// Track the outbound row count between result events. When the agent uses
// send_message (or any MCP tool that writes to messages_out) during a turn,
// the count grows. We pass that signal to dispatchResultText so it can tell
// the difference between "agent wrote text meant as the reply" (send the
// scratchpad) and "agent did explicit tool sends AND then emitted a trailing
// status line" (don't echo the status line back to the channel).
//
// Reset after each result dispatch so subsequent turns in the same query
// (follow-up messages pushed into the stream) are evaluated independently.
let outboundAtLastResult = getOutboundCount();
// Concurrent polling: push follow-ups into the active query as they arrive.
// We do NOT force-end the stream on silence — keeping the query open is
// strictly cheaper than close+reopen (no cold prompt cache, no reconnect).
@@ -323,7 +334,9 @@ async function processQuery(query: AgentQuery, routing: RoutingContext): Promise
if (event.type === 'init') {
queryContinuation = event.continuation;
} else if (event.type === 'result' && event.text) {
dispatchResultText(event.text, routing);
const hasExplicitSends = getOutboundCount() > outboundAtLastResult;
dispatchResultText(event.text, routing, hasExplicitSends);
outboundAtLastResult = getOutboundCount();
}
}
} finally {
@@ -363,7 +376,7 @@ function handleEvent(event: ProviderEvent, _routing: RoutingContext): void {
* This preserves the simple case of one user on one channel — the agent
* doesn't need to know about wrapping syntax at all.
*/
function dispatchResultText(text: string, routing: RoutingContext): void {
function dispatchResultText(text: string, routing: RoutingContext, hasExplicitSends: boolean): void {
const MESSAGE_RE = /<message\s+to="([^"]+)"\s*>([\s\S]*?)<\/message>/g;
let match: RegExpExecArray | null;
@@ -397,7 +410,15 @@ function dispatchResultText(text: string, routing: RoutingContext): void {
// Single-destination shortcut: the agent wrote plain text — send to
// the session's originating channel (from session_routing) if available,
// otherwise fall back to the single destination.
if (sent === 0 && scratchpad) {
//
// If the agent already sent messages explicitly this turn (via send_message
// or another MCP tool that writes to outbound), treat trailing plain text as
// a status/summary line and DO NOT echo it back to the channel. Without this
// guard, task-driven flows like the onboarding /welcome cause duplicate
// delivery: the skill uses `send_message` to greet the new user, then the
// model emits "Welcome message sent." which used to be dispatched as a
// second chat message to the same recipient.
if (sent === 0 && scratchpad && !hasExplicitSends) {
if (routing.channelType && routing.platformId) {
// Reply to the channel/thread the message came from
writeMessageOut({
@@ -422,11 +443,15 @@ function dispatchResultText(text: string, routing: RoutingContext): void {
log(`[scratchpad] ${scratchpad.slice(0, 500)}${scratchpad.length > 500 ? '…' : ''}`);
}
if (sent === 0 && text.trim()) {
if (sent === 0 && text.trim() && !hasExplicitSends) {
log(`WARNING: agent output had no <message to="..."> blocks — nothing was sent`);
}
}
function getOutboundCount(): number {
return (getOutboundDb().prepare('SELECT COUNT(*) AS c FROM messages_out').get() as { c: number }).c;
}
function sendToDestination(dest: DestinationEntry, body: string, routing: RoutingContext): void {
const platformId = dest.type === 'channel' ? dest.platformId! : dest.agentGroupId!;
const channelType = dest.type === 'channel' ? dest.channelType! : 'agent';