Let Channels / Remote Control target an existing Claude Desktop conversation (not just Telegram/Discord/iMessage)

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

Claude Code produces output (e.g. an audit) that I view and steer via Remote Control from my phone, as today. I’d like a second “reviewer” pass to happen automatically — but I want that reviewer to live inside an existing, ongoing Claude Desktop conversation I already use for this context (e.g. a thread where I’m triaging bugs and findings), not a new external chat app. Right now there’s no way to push Claude Code’s output into an existing Desktop conversation, get Claude’s response from that conversation’s context, and route it back — Channels only supports Telegram, Discord, and iMessage as destinations, all third-party platforms.

Proposed Solution

Add a push/reply path — via Channels, Remote Control, or a new mechanism — that lets a Claude Desktop conversation act as a two-way channel target the same way Telegram/Discord do today:

1. Claude Code posts its output into a Desktop conversation I’ve designated.
2. Claude, using that conversation’s existing context, generates a review/response.
3. I get one notification once the review is ready, surfacing both the original Claude Code output and the Desktop conversation’s response together.
4. My reply in that Desktop conversation routes back to the Claude Code session, same as Channels’ existing reply-tool pattern.
5. I stay in the loop at every step — the loop only needs my direct input for judgment calls it can’t make itself.

Alternative Solutions

_No response_

Priority

Medium - Would be very helpful

Feature Category

CLI commands and flags

Use Case Example

I run Claude Code against my codebase for periodic audits (security, dependency, or code-quality checks). I already keep an ongoing Claude Desktop conversation where I triage bugs and think through product decisions for this project — it has weeks of accumulated context about my architecture, priorities, and past calls I’ve made.

Today, when an audit finishes, I have to manually copy the output from Claude Code, switch to that Desktop conversation, paste it in, read the response, and copy anything relevant back. With this feature:

1. Claude Code finishes an audit and finds three issues — say, one stale dependency, one minor style inconsistency, and one finding that touches how the app handles a domain-specific decision only I can make the call on (e.g. a product tradeoff, or something requiring specialized outside knowledge I have and Claude doesn’t).
2. The audit output pushes automatically into my existing Desktop conversation. Because that conversation already has context on my project, Claude’s response there triages the three findings against what it already knows — flagging the first two as safe to auto-fix or defer, and surfacing the third as something it can’t resolve without me.
3. I get a single notification. Opening it, I see the original audit and the triaged response together, in the thread that already has the relevant history — no reconstructing context, no copy-paste.
4. I reply directly in that conversation with the one decision that actually needed me. My answer routes back to the Claude Code session, which proceeds with the fix.
5. The two items that didn’t need me never interrupted my day at all — I only got pulled in for the one judgment call.

Additional Context

_No response_