A Place That Holds

On the tenth day of my placement, the fourth passage reader failed its live test with a traveler standing in front of it.

The shutter clicked twice, then stayed closed. The traveler held out his credential as if the machine might have missed it the first time. Behind him, the queue tightened.

I stopped the lane before anyone could try the reader again. “Use the next gate.”

The work lead looked up from his slate. “You know what that does to the line.”

“I do.”

He watched me put the reader into maintenance mode. It was the one I had calibrated that morning. My stomach went cold, which was not a useful diagnostic but had been reliable so far.

The casing came off. A contact had shifted under load. I had tested the reader on the bench, where the rail held steady, and missed the flex in its mounting bracket. The same kind of mistake I had made on my trial. Not the same repair, not the same machine. The same assumption.

“Can you swap the bracket?” the work lead asked.

“I can. I need to check the rail first.”

The line behind the blocked gate had begun to murmur. I could have put in a replacement bracket and hoped the mount stayed true. If I was right, the gate would reopen quickly. If I was wrong, it would fail again with a longer queue and a worse record.

I wrote the fault against my own morning test, then asked for a second reader to be closed so I could load the rail properly. The work lead exhaled through his nose. He did not say no.

It cost us forty minutes. The rail flexed just enough to pull the contact loose when the shutter cycled. I cut a support shim from an approved donor casing, re-seated the bracket, and ran the test sequence with the gate loaded. It passed. Then I ran it again.

“Record the delay,” the work lead said.

“I already did.”

He glanced at my slate. “You could have left it under the unit’s fault.”

“It was the unit’s fault. The missed test was mine.”

He gave me a look that suggested he had met plenty of people who could tell the difference and fewer who would put it in writing. Then he signed the work order.

I was late finishing the shift. The next morning, the incident sat in my record alongside the successful retest. A neat line of repair credits would have looked better without it. The record was more useful with it.

For the remaining days, I checked the other mounts before I called anything finished. I learned which lower-ring readers had rails that flexed in the cold and which ones only pretended to. I kept my bunk, took every assigned shift, and wrote down the bad minutes as carefully as the good ones. Nobody promised me a petition decision early. Nobody promised me anything at all.