The spine operation ran at three in the morning.
Carrow took six nodes in ninety-four minutes.
Not simultaneous — sequenced, the timing calculated to prevent any single node from sending a warning signal before the next one was addressed. Reyes had mapped the inter-node communication intervals from the activation pattern data, and Carrow had built the sequencing around those intervals with the specific competence of someone who had been planning this operation for two weeks and knew the material.
Adrian received the confirmation at four forty.
Six nodes down.
The coalition's coordination infrastructure was resolved.
He held the confirmation for a moment and ran the remaining picture: the three organizations without coordination capacity, the seven-location attack running on its own momentum but losing cohesion as the night progressed, the war's structure shifting from the coalition model back to three separate organizations that hated each other.
The coalition was functionally over.
The war wasn't.
What remained was the three organizations operating independently, which meant three separate operational pictures and three separate response requirements. The Petrov Bratva had the most depth — the longest institutional history, the most organizational resilience. The Ortega remnant had the most aggression. The eastern districts organization had the most to lose and was therefore the most unpredictable.
All three knew the coalition had failed.
All three knew this was the last operational phase.
Cassian understood this the way he understood strategic pictures — from the shape first, then the components.
"They have nothing left to lose," he said at the morning briefing. "That makes the next two weeks more dangerous than anything that's come before."
He said it in the flat accurate register.
"The coalition was holding them to a coordinated methodology," Adrian said. "Without coordination, each organization will revert to its own approach. We need three separate operational pictures."
"We have them," Carrow said.
"Reyes," Cassian said.
"The Petrov picture is the most complete," Reyes said. "The Ortega remnant's structure has gaps — we lost two nodes in the eastern approach collapse and haven't rebuilt coverage. The eastern districts organization has been quiet since the coalition's spine went down, which could be regrouping or could be something else."
"Map the quiet," Cassian said.
"Already running," Reyes said.
The briefing continued.
Cassian ran it with the quality that had shifted since the seven-location night — the patience still present but the tempo different, the decisions coming at a different speed. He was pushing the pace. Not urgency in the reactive sense — the deliberate acceleration of someone who had assessed the window and knew it was closing.
Adrian ran beside him.
This was the phrase that had settled into the war council's language in the past week — the specific configuration of how the two of them operated when the operational picture required both the strategic layer and the field layer simultaneously. Cassian held the strategic picture and pushed the pace. Adrian held the field operations and executed with the precision that had been the Wiper's professional signature for six years.
Together they moved through the war's final phase like two predators that had been hunting in parallel and had found the same quarry.
The city felt it.
The operations ran every night.
Not all of them were combat operations — the war in its final phase was as much about the organizational picture as the physical one. Petrov's financial infrastructure, dismantled through three targeted operations that Reyes had designed and Adrian had executed. The Ortega remnant's leadership communication, disrupted through a combination of the network's coverage and two direct engagements with Ortega field commanders that had the specific quality of encounters between people who knew they were running out of time.
The eastern districts organization remained quiet.
This was the variable that Adrian kept returning to — the quiet that was either regrouping or something else. He'd run the assessment twice and arrived at something else both times and couldn't confirm it.
He raised it with Cassian on day four.
"The eastern districts quiet isn't passive," he said. "Passive quiet looks different. This has the quality of active restraint — an organization that's choosing to hold rather than an organization that's exhausted."
"They're waiting," Cassian said.
"For something specific," Adrian said. "Not for the Petrov or the Ortega situation to resolve. For something in the Syndicate's picture."
Cassian held the map.
"Sable," he said.
"The message said come face me if you want this war to end," Adrian said. "If the eastern districts organization is connected to Sable rather than to the coalition — if they've been Sable's organizational base rather than coalition partners—"
"They're waiting for Sable to deliver," Cassian said.
"They're waiting to see if the message produces the outcome," Adrian said. "If I respond to the message and the response goes in Sable's favor, they have an opening. If I don't respond, they eventually act independently."
"Either way they move," Cassian said.
"Yes," Adrian said. "The question is when and from what position."
He held the map.
"I haven't found the third penetration yet," he said.
Cassian looked at him.
"Reyes has the investigation running," Adrian said. "The access data from the safe house rotation and the pre-modification approach route narrows it to eleven people. Eleven people had access to both data points."
"Eleven is manageable," Cassian said.
"I'm working through it," Adrian said. "When I have the penetration, I have Sable's location. When I have the location—"
"You tell me," Cassian said.
"Yes," Adrian said.
Day six.
The Ortega remnant's field commander made the mistake of using a predictable approach route.
Adrian knew the route from the Ortega intelligence thread. He'd flagged it in the war council three days before as the probable route for any Ortega operation in the middle district. The flag had been noted and the position had been prepared.
The field commander arrived at the prepared position.
The engagement lasted four minutes.
The Ortega remnant's field-level command was now in significant disarray, which would have implications for the Ortega organization's operational capacity for a minimum of two weeks and possibly longer depending on whether they had succession depth.
Adrian assessed the engagement on the walk back to the vehicle.
He reported to Cassian via the channel.
Cassian said: "The Petrov operation tonight is still on."
"Yes," Adrian said.
"Doran has the team."
"I'll be there," Adrian said.
The Petrov operation ran at midnight. It was the most complex of the war's engagements — the Petrov Bratva's organizational depth meant they had redundancy in their infrastructure, and the operation needed to address three separate financial positions simultaneously to produce the pressure the strategic picture required.
He ran it with Doran and eight people.
It took twenty-two minutes.
The Petrov Bratva's financial infrastructure had three fewer functional components at the end of it.
He walked out of the last position at twelve forty and had the specific quality of someone who had been running at full capacity for six consecutive days and whose body was implementing that assessment.
He noted the assessment.
He filed it.
He continued.
The raid on day eight was the Petrov secondary command position.
Not the primary — the primary was defended with the depth of an organization that knew it was running out of options and had concentrated its remaining resources at the core. The secondary had less defense and more operational value: the communication infrastructure for the Bratva's remaining field operations, the thing that needed to be removed before the primary could be addressed.
He had the secondary's layout from Reyes's intelligence — the building, the access, the guard configuration, the communication equipment's location.
The approach was clean for three minutes.
Then it wasn't.
He knew it at the thirty-second mark of the fourth minute.
He knew it the way he'd known things in engagements for six years — not from a specific visual or audio confirmation, from the quality of the space changing. The building had too many people in it. Not the six that Reyes's intelligence had confirmed. More.
He was already assessing the revision when the first position activated.
Not from the guard configuration Reyes had mapped.
From positions that had been placed after the intelligence was gathered.
He ran the geometry in two seconds: eight positions, maybe ten, covering all the approaches he'd entered from and all the exits he'd identified.
Not a coincidence.
The positions were specific to his approach.
Someone had known his approach before he'd arrived.
The third penetration.
He was in the trap before the assessment had finished.
He moved.
He moved the way he moved when the geometry had closed — not toward the exits the trap was covering, toward the trap's structure itself. The specific counter-intuitive logic of a prepared position: the covering positions focused outward, which meant there was a gap at the trap's center. The gap wasn't safe but it was less covered than the perimeter.
He reached the center position's structural cover in eight seconds.
He had his phone.
He had the channel.
He had Cassian's condition — when you find the location, tell me before you move.
He'd found the location.
He was already in it.
He pressed the channel.
"Secondary command position," he said. "It's a trap. Eight to ten positions, covering all mapped approaches. Someone gave them my approach geometry."
A pause.
Three seconds.
Cassian's voice: "Hold the center."
"Holding," Adrian said.
"Doran is four minutes out," Cassian said.
"The trap was built for one person," Adrian said. "Ten positions for one target. If Doran arrives with a full team—"
"The trap becomes surrounded," Cassian said. "I know."
"Four minutes," Adrian said.
"Three now," Cassian said.
He held the channel.
He held the center position's structural cover.
He looked at the ten positions in the building's available light and ran the assessment of three minutes against the coverage and what three minutes required him to do.
He started the assessment.
He began.