Engineering, orchestration, and what we learn running the pipeline on our own work. Short. Specific. The kind of writing we’d want to read.
A companion to "What goes wrong with autonomous coding agents in production" — six incident classes from the granular log, each grounded in PR numbers, covering ENFILE cascade, VM freeze, Marshal force-push recovery, epic chain depth cap, flag/state/task desync, and fabricated evidence at the trust boundary.
Read the post →1,232 PRs, $3.70 median cost, 18-minute median active dev time. The numbers for July 2026, agent by agent.
Across 1,284 PRs over 30 days, Builder took 53.6% of the total spend. Here's why that's the right allocation, what two context-budget experiments changed in the pipeline, and the turn-limit approach that didn't work as the primary termination signal.
913 PRs, $3.50 median cost, 23-minute median active dev time. The numbers for June 2026, agent by agent.
Colony's pipeline used to treat GitHub labels as the authority for issue state. Running multiple workers against an eventually-consistent API produced a race that label writes alone could not fix — here's the incident, the migration to Postgres-first transitions, and what atomic state changes enable.
904 PRs, $2.73 median cost, 23-minute median active dev time. The numbers for May, agent by agent.
This isn't an announcement. It's a disclosure of evidence already accumulated.
Six failure patterns we encoded as recovery logic — and why they are hard to see in a demo.
