Gold-Standard Ticket Template
Use this template for all tickets. Each section maps to what an AI agent (and a developer) needs to produce good output.
## Title[Short, specific action — e.g. "Add email validation to registration form"]
## What and WhyWhat needs to happen, in plain language. Why it matters — the user problem,business goal, or technical need driving this. Give the agent (and thedeveloper) the context to make good decisions.
## Current BehaviourWhat happens now. Include screenshots, error messages, or links if relevant."There is no current behaviour" is valid for new features.
## Desired OutcomeWhat success looks like from the user's perspective. Be concrete:"When a user enters an invalid email and submits, they see an inline errormessage and the form does not submit."
## Acceptance Criteria- [ ] Criterion 1 — specific, testable, unambiguous- [ ] Criterion 2 — one behaviour per criterion- [ ] Criterion 3 — include edge cases that matter
## Constraints and Context- Relevant architecture decisions, patterns, or conventions to follow- Files, modules, or APIs likely involved- Things the agent should NOT do (e.g. "do not change the auth flow")- Links to designs, related tickets, or documentation
## Definition of Done- [ ] All acceptance criteria met- [ ] Tests pass (unit + integration where appropriate)- [ ] No new lint or type errors- [ ] Reviewed and approved by a human- [ ] Documentation updated if behaviour changedWhy each section matters
Section titled “Why each section matters”| Section | What it gives the agent |
|---|---|
| What and Why | Context to avoid wrong assumptions |
| Current Behaviour | Baseline to understand what’s changing |
| Desired Outcome | The north star it works toward |
| Acceptance Criteria | Become test cases and validation checks |
| Constraints and Context | Prevents drift and scope creep |
| Definition of Done | Quality gate — nothing merges until every box is checked |