Hand a feedback request to your coding agent

Description: Assign a feedback post to your own coding agent — Claude Code, Cursor, GitHub Copilot, or one you built yourself — and get an implementation plan back, attached to the post as a document. You connect your own agent account, so the work runs on the agent you already pay for. Sleekplan hands over the request and receives the thinking back. It never touches your repository — your agent already has your code through your own connection. Three principles we're designing to: A plan first, code second. The default result is an implementation plan you can read, not a pull request you have to review. Opening a PR would be opt-in. A person stays responsible. The agent is a delegate, not the new owner. It can draft, comment, and move internal states — only a person publishes a public status or promises anything to your users. One request stays one request. Delegating doesn't split the post into sub-items or start a second queue. Problem: Getting from "customers keep asking for this" to "here's what it would take" is manual work. You re-type the request into a ticket or a prompt, the agent's answer lands in a tool nobody else opens, and the thinking gets separated from the request that prompted it. It also means effort is guesswork at the moment you most need it. You can see how much demand a request has, but not what it would cost to build — so prioritisation leans on the half of the picture that's easy to measure. Benefit: Scope a request without leaving the board, and keep the plan next to the feedback that prompted it. Get an effort signal to weigh against demand, instead of ranking by votes alone. Keep using the agent your team already chose, on your existing plan, with your code staying where it is. We'd like your input: Which coding agent does your team use today? Is an implementation plan useful on its own, or would you only want a pull request? Who on your team should be allowed to hand work to an agent?