That evening Maua requested permission for a simple field observation. She did not bring Nuru. She did not want the dog's behavior changed by the child he knew best, and she did not want her daughter turned into evidence.
Dr. Wanjiku allowed Kito to be walked to the kiosk by a volunteer on a long lead and then observed in a controlled section after a risk check. Silas came to watch what he called "the routine story."
"It isn't a story," Maua said. "It is an observation."
"Residents won't hear the difference."
"Then we make it clear."
At 4:02 the first pupils left school. Kito lay under a tree, nose to the ground. At 4:06, when a cluster of uniforms passed, he lifted his head. He did not follow immediately. He moved only after one child dropped a food wrapper near the kiosk bin.
Kito stopped at the kiosk because Mugo poured water outside. He drank. The children continued. After several seconds he moved behind them. In the narrow lane he stopped to smell a tyre and fell far back, then increased pace toward a group of uniforms.
"He follows the children," Silas said.
"Yes," Maua replied. "We still need to know what triggers it and what interrupts it."
At the estate crossing two children entered. Kito stayed outside. Another group appeared behind him. The dog turned and went back toward the kiosk with the second group instead of remaining with the first.
Dr. Wanjiku wrote: movement cue, familiar time, food association, group.
They repeated the observation without a child in the lead group. A volunteer carried a school-type backpack. Kito followed briefly, then abandoned the route when he heard a biscuit packet near Mugo's kiosk.
"Exactly," Silas said. "Unreliable."
"Yes."
He looked surprised.
"But the unreliability of the dog does not erase the unreliability of our custody system," Maua said. "You can remove Kito and the child can still be stranded."
"Zero-stray removes one variable."
"Fine. Remove a variable if your policy justifies it. Do not call that the entire safety plan."
They reached the estate gate. Joseph confirmed the pedestrian side gate was closing at 4:20 on sweep days.
"What happens to children arriving at 4:20?" Maua asked.
"The vehicle gate remains open."
"You expect children on foot to use the vehicle lane?"
"A guard escorts them."
"Is that recorded?"
Silas said, "We cannot write every common-sense step."
"That is what Daniel and I said about handoff. Common sense was an assumption."
At 4:32 Daniel arrived, his ordinary Tuesday time. He looked at the gate and then back down the route.
"Would Nuru already be inside or waiting outside?"
"Depends on the minute," Maua said.
Dr. Wanjiku's summary was deliberately boring: route-associated following under familiar cues; not reliable escort behavior; no aggression observed in the controlled session; continued supervision and foster care recommended.
Maua asked that the school receive the summary too, because the original incident had happened in their corridor. The principal agreed to place it in the safety file without labeling Kito either mascot or threat.
At home, Maua explained the result to Nuru.
"Kito follows groups, smells, time, food. He is not deciding who needs protection."
"I know."
Maua slowed. "You knew?"
"I'm nine, not stupid. I said he comes with us."
The adults had been the ones adding purpose.
Silas sent a message before sunset: **Stray-control operation begins 06:00 tomorrow. Pedestrian side-gate closure protocol remains active.**
Kito might be removed in the morning.
The gate would still close early.