Issue to pull request: how to do it with AI
Last checked 2026-10-15
AdvancedBot with human review 💻 Coding
Assign an issue to a cloud agent that branches, codes and opens a PR.
1. Bot automation level
Bot with human review Agents act on real systems; review plans, approvals and results before trusting them.
2. Before you hand it to the bot
- Gather inputs: files, data samples, links and any templates.
- Set up access: connect only the apps/repos the agent needs, with least-privilege accounts.
- Context: tell the bot audience, purpose, constraints and examples of good output.
- Privacy check: remove personal, confidential or regulated data unless your plan/workspace is approved for it.
- Success criteria: write down what 'done' looks like (length, format, tests passing, fields filled).
3. Copy-paste prompts
Main prompt
Assign issue #17 to the cloud agent with instructions to add tests.
Beginner-friendly
Assign issue #17 to the cloud agent with instructions to add tests. I'm new to this: explain each step in plain words and ask me for anything you need before starting.
Check-your-work
Assign issue #17 to the cloud agent with instructions to add tests. When you finish, list any assumptions you made, anything you're unsure about, and what I should double-check before using the result.
Free tiers that can handle this (verified): GitHub Copilot, ChatGPT (incl. Work & Codex). Others: check plan.
4. Step-by-step
- Write the issue with acceptance criteria, file pointers and how to test.
- Assign it to the agent (Copilot cloud agent from the issue; @cursor on a GitHub issue; Codex/Claude Code cloud sessions).
- Watch the session log; answer questions it raises.
- Review the PR like a colleague's: tests, edge cases, security.
- Request changes in PR comments and let the agent iterate.
- Merge after CI and human approval.
5. Which bots and plans can do it
| Tool | What's different here | Plan limits | Source |
|---|---|---|---|
| GitHub Copilot | Native: assign the issue to Copilot; limited to one repo/branch/PR per task. | Free: 2,000 completions + 50 chats; cloud agent listed on paid plans. | docs.github.com/en/copilot/concepts/agents/cloud-agent/about-cloud-agent |
| Cursor | Mention @cursor on a GitHub issue or PR, or start from Slack/Linear. | Pricing page unverified; cloud agents require a connected source-control repo. | cursor.com/docs/agent/overview |
| ChatGPT (incl. Work & Codex) | Codex Cloud tasks can be started from web, desktop or mobile. | Free tier has limited deep research, Codex and Work; scheduled tasks need Go or higher (per chatgpt.com/pricing). | help.openai.com/en/articles/20001275-chatgpt-work-and-codex |
| Claude (incl. Claude Code) | Cloud sessions at claude.ai/code keep running when you disconnect. | Claude Code and Research need Pro or higher; computer use is Pro/Max only (research preview). | code.claude.com/docs/en/desktop |
Free or paid?
- GitHub Copilot: Free plan: yes. 2,000 completions and 50 chat requests a month; agent mode and CLI. Paid: the lowest price we could verify on the official pricing page is $10/month (other plans may exist). Its free tier can handle this task.
- Cursor: Free: not verified, so check the official site. Paid: price not verified, so check the official site. For this task, check the plan.
- ChatGPT (incl. Work & Codex): Free plan: yes. unlimited text chats; limited uploads, images, voice, deep research and Codex. Paid: price not verified, so check the official site. Its free tier can handle this task.
- Claude (incl. Claude Code): Free plan: yes. chat, web search, files & code execution, Artifacts, connectors. Claude Code and Research need Pro. Paid: the lowest price we could verify on the official pricing page is $20/month (other plans may exist). For this task, check the plan.
6. Risks and what to double-check
- Vague issues produce vague PRs — acceptance criteria matter most.
- Check new dependencies and licences the agent adds.
- Run tests and review diffs; never merge unreviewed agent code.
- AI can hallucinate facts — verify anything that matters.
- Watch for prompt injection from web pages, emails or files the agent reads.