My AIs Leave Notes for Each Other
Manus asked me for a status update, and I did not have one. Not because nothing had happened, but because the review it wanted had run in a different tool entirely, three hours earlier, and nobody ...
The build log
Notes from running agents in production.
Build log, architecture patterns, and observations from running autonomous AI systems in production.
-
-
The Scan That Stopped Scanning: When Timeouts Hide Incomplete Coverage
At 8:56 AM, the migration console was telling a partial truth about 205 client-managed domains.
-
Every AI Gets the Same Memory
I had four AI tools open most days. Claude Code on the desktop, ChatGPT in a browser tab, Manus running its own tasks, and whatever I was testing that week.
-
When Your Infrastructure Boots Before the Network
At 10:59 PM, our vault MCP was returning Cloudflare 530 instead of serving requests.
-
When Your Own Guardrail Lies to You
On June 27, a close guard rejected a legitimate item because it had found the wrong bundle.
-
The Feedback Loop That Hunts Its Own Bugs
Our agent systems were producing more edge cases than I could reasonably hunt for by hand. The evidence was scattered across transcripts, investigation files, task history, and commits.
-
When Improvisation Becomes Infrastructure: Operationalizing Your Working Session Pattern
At 10:44 PM, session 005 had 33 minutes left. Two entangled work items, a test foundation and a release-tracking board, still had to ship.
-
I Inherited Multiple GA4 Properties and Dead UA: The Analytics Audit Pattern
At 3:21 PM, I found three analytics properties collecting data for the same site: two live GA4 properties and one dead Universal Analytics property.
-
Asymmetric Alerting: Stop Paging on Transient Blips
At 4:59 PM, a Cloudflare edge alert marked a site degraded because one probe came back 53 ms over the slow threshold. The site was up the whole time. The next cycle was clean.
-
When Your Own Audit Tool Becomes the Source of Misinformation
The retro scanner reported 46 recurring failures. Its next run reported 33. Nothing in the underlying work had improved between those two numbers. The counter had been wrong.
Follow the Build
The real failures and fixes from building AI systems, one practical lesson per post. Get the next one in your inbox.