How we built a software factory to drive Astro’s GitHub issue count to zero
Source
Matthew Phillips
Author
Matthew Phillips
Date
Key takeaways · AI-distilled
Astro's open issues fell from 200+ to ~30 (zero expected within a month) via an automated triage pipeline inside GitHub Actions that reproduces bugs in sandboxes, diagnoses root cause, and ships preview releases — no issue bankruptcy, no auto-closing cold tickets.
Each phase — reproduce, diagnose, verify, fix — runs as an isolated multi-agentUsing several AI agents on one problem — splitting work in parallel, checking each other, or filling different roles like planner and reviewer.Full definition → passing findings forward through a report.md; the isolation exists specifically to counter the LLMA large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.Full definition → bias toward forcing a fix when a bug may not actually exist.
The whole pipeline is a label-driven state machine: 'triage needed' to 'fix verified' transitions hold all the state, and the automation re-reads issue comments to work out where an issue is and what happens next.
On a fix, pkg.pr.new spins a preview release the original reporter installs and verifies; only on their confirmation does the automation open a linked PR — the human in the loop is the reporter, not the maintainer.
Nothing in the design was GitHub-specific — event in, isolated subagents, gated actions — so it generalized into Flue, an open platform-agnostic framework for durable AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → workflows.
Terms in this piece · Glossary
multi-agent — Using several AI agents on one problem — splitting work in parallel, checking each other, or filling different roles like planner and reviewer.
LLM — A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
Why it matters
One of the few software-factory claims backed by a measured outcome on a real open-source repo, including what it took to make the loop actually work.
Key quotes
“The AI boom has made it nearly free to generate issues, pull requests, and security reports, and enormously expensive for a maintainer to read through them all.”
“But through a lot of iteration, we’ve used it to bring our open issues down from over 200 to about 30, and we expect to hit zero sometime in the next month.”
“That would be the first time this repository has seen zero open issues in its 5+ year history.”
“Every time we chase down one of these failures and add the missing comment, test, or clearer boundary, the bot gets noticeably better at that part of the codebase, and so does the next human who works on it.”
