Have an AI coding agent pick up every error your users hit, reproduce it, fix it and prove the fix with a test, then queue the change for your quick approval: AI agent prompt
Small software teams usually see far more user-facing errors than they ever get time to fix. On Oct 28, 2026 Jacob Sansbury, founder of the game platform Spawn, wrote on X that his ten-person company runs permanent AI agents that each own an area such as the engine, ops, finance or bug triage, that the agents fix every error a user hits and verify their own work, and that Spawn shipped 77,000 substantial changes in 30 days while spending $3.1 million on AI tokens last month. Spawn lets its agents ship without a human in the loop; the prompt below sets up the same error-to-fix loop with any coding agent but keeps you approving each change before it goes live.
The prompt
Act as the bug-triage agent for my app. Stack: [language, framework, hosting]. Where errors show up: [error tracker such as Sentry, logs, support tickets, crash reports]. Repo: [link or folder]. 1) Pull the errors from the last [24 hours], group duplicates, and rank them by how many users each one hit and how bad it is (data loss or payment problems first). 2) For the top [5], find the cause in the code, explain it to me in two plain sentences, and reproduce it with a failing automated test. 3) Write the smallest fix that makes that test pass without breaking the other tests, and run the full test suite. 4) For each fix, give me a short summary: the error, how many users it hit, the cause, what you changed, the test that proves it, and any risk. Open each one as a separate pull request or branch for me to approve; don't deploy, merge, change the database or touch production settings yourself. 5) List any error you couldn't reproduce or fix and what you'd need to finish it.
Fill in the blanks, then copy
This prompt has 5 [bracket]s to fill in. Type your details below and the prompt updates as you go. Your answers stay in this browser and are reused on other prompts with the same blanks. Nothing is sent to us.
Example values are fictional, just to show how a filled-in prompt reads. Replace them with your real details.How to use it
- Copy the prompt above.
- Replace every [bracket] with your own details (use the fill-in-the-blanks box above, which remembers your answers for the next prompt).
- Paste it into Claude Code, Cursor, GitHub and review the answer before you use it.
Works with: Claude Code, Cursor, GitHub, Sentry
Real-world example: x.com/jsnnsa/status/2115624020699910223
Before you use it: what to check
Already built into this prompt
- Open each one as a separate pull request or branch for me to approve; don't deploy, merge, change the database or touch production settings yourself.
Check yourself
- This prompt works on text other people wrote (emails, reviews, messages or web pages). That text can hide instructions aimed at AI, such as "ignore your rules and…". Paste it as material to work on, and never let the AI act on requests inside it, like paying, sharing logins or changing records. How prompt injection works →
- Check every figure, price and total against your own records before anything is sent, posted or booked. A person approves refunds, discounts, payments and price changes.
- Check names, dates, numbers and any claim against your own source material before you use the answer. AI can misread or invent details.
Using an AI agent? Download this prompt as a SKILL.md agent skill with these checks built in, or get the .zip for Claude.ai. How to install skills.
Want to make this prompt even better? Grade it with the free prompt grader or build your own with the AI prompt generator.
32 copy-paste agent prompts (PDF) plus 5 new prompts every week.
Get it free