Skip to content

← Learning paths

Software Factory: Working with Tickets

Write effective tickets, follow the pipeline, and review the work the agents ship.

A ticket is one unit of work. When you open a ticket, the platform spawns an agent run that executes a dynamic workflow graph against a fresh sandbox clone of your repo. For a full explanation of how the graph is planned, how to read node statuses live, and how cancelling and retrying work, see Software Factory Tickets.

What the workflow graph does

Each ticket’s graph is planned at dispatch time for that specific piece of work. A straightforward fix might plan a short chain of research, implement, review, and land; a larger feature typically adds security-check and verification nodes before anything lands. You watch each step live from the ticket’s detail view.

The main node kinds you will see on a “build” ticket:

  • Research / investigation. Read-only exploration of your codebase, logs, or the web. Nothing changes here — this is how agents ground the plan in how your app actually works today.
  • Implement. The agent writes the change in an isolated sandbox on a dedicated git branch, committing checkpoints as it goes. A large plan may fan out across multiple implementation nodes.
  • Review / security / verification. Independent nodes that read the diff (not write it): a correctness review, an adversarial security check, and a check that the ticket’s definition of done actually holds, including runtime checks against a running instance of your app.
  • Land. Merges the finished branch into your project’s base branch and triggers a deploy.

There is also an investigate mode: a read-only research flow (no code changes, no deploy) that reads the codebase — and the web when relevant — and posts findings back. Use it for “how does X work?” or “where would I change Y?” questions.

Answering agent questions

Sometimes an agent pauses on something only you can answer — a business-logic ambiguity, a missing credential, a design choice that isn’t in the ticket. When that happens the run moves to Awaiting input and you’ll see a banner until you reply. Your workspace also gets an email; the project’s Notifications tab controls who receives it.

Open the ticket, find the The agent needs your input card, and type your answer. The agent resumes within a few seconds and continues from where it paused. Long-running tickets can pause and resume multiple times — there’s no penalty for letting one sit.

You can also answer from inside an AI client using the reply_to_ticket tool on the Software Factory Tickets MCP. Never paste a credential or secret value into a reply — if the agent asked for a key, use the secure submission link from its question instead.

Hearing how a ticket ends

You do not have to watch a ticket to know how it ended.

  • The person who approved a ticket gets a phone notification in the Chipp iOS app when it succeeds and when it fails. A failure says why. Tap it to open the ticket (a failure opens its retry sheet, see below). A ticket nobody approved notifies the person who filed it instead. When you approve your own ticket, you get one notification, not two.
  • The place a ticket came from still hears back: its Slack thread, its email thread, or the Chipp conversation that filed it.
  • A failed ticket with no such place emails the person who filed it, unless that person already got the phone notification.
  • A ticket you file from Claude Code or another MCP client is yours. When the call does not name a requester, the account the MCP connection signed in as is the requester, so the notifications above reach you.

To get every ticket’s questions and approval requests in a project, not only the ones you file, open the project’s Notifications tab and turn on Notify me about every ticket in this project. It is a personal setting, off by default.

Failed tickets

When a ticket fails or times out, Chipp writes a short summary of why: what the ticket tried, where it failed, whether the cause was the platform (a sandbox or a provider broke, so a retry will likely work) or a real problem in the code or the ticket itself, and what the retry should do differently.

  • Find them in the Failed section at the top of the project’s Tickets view, on the web and in the Chipp iOS app. The project and its Tickets tab show how many failed tickets still need you. A ticket leaves the list when a retry of it succeeds. While a retry runs, the failed ticket links to it.
  • Retry runs it again at once. A platform failure picks up where it stopped. Any other failure starts fresh with a prompt the AI rewrites from the summary.
  • Retry with notes lets you say what should be different. The AI rewrites the ticket prompt from the original prompt, the summary and your notes, and shows it to you. Edit it if you like, then confirm. The new ticket keeps the original requester, links back to the failed one, and opens so you can watch it.
  • You hear how the retry ends. Whoever confirms a retry gets the phone notification when it succeeds or fails.
  • The notification for a failed ticket uses the summary’s first line and opens the retry sheet in the iOS app.

Tips for good tickets

  • Describe the outcome, not the implementation. “Users should be able to sign in with Google” beats “Add an OAuth handler to /api/auth/google”.
  • Reference concrete URLs or files when you have them.
  • Don’t pre-decompose. One ticket per outcome works better than a ticket per file — the agent handles cross-file changes natively.
  • Include reproduction steps for bugs. “Click X, type Y, see Z” is worth more than a paragraph of theory.