After Action

The secure link opened at 1400 station time, eight hours after we'd come back through the drainage channel with Ollen's arm in a field stabilizer and Maret in a body bag.

Vasht's voice was level, the way it always was, which meant the debrief would run exactly as long as it needed to and not one second longer. Greel had his arm in the proper stabilizer now. Denavi had her terrain assessment queued. I had the mission sheet open on the equipment case next to me, the notations visible, because I'd learned in the first two debriefs that Vasht asked questions in the order they mattered and I needed the data ready rather than reconstructed from memory.

"Forward unit, this is a formal after-action review of the Caleth Corridor node operation. Timestamp 1400, recording active." A short pause, the kind she used to let everyone settle. "I'll cover the operational sequence first, then individual performance assessments. Greel, confirm unit status."

Greel gave the status in his precise, unhurried way: two KIA, two wounded, one of those wounds operationally limiting. He named Maret. He named Pvt. Cors. He gave Ollen's injury status and Greel's own arm injury without any emphasis on the latter, which told me he'd been delivering this kind of report for a long time and had learned to flatten his own name into the list the way you'd flatten any other data point.

"Confirmed," Vasht said. "Derek."

"Here."

"Your initial position at point of first contact: northwest approach, covering the primary access."

"Yes."

"At the moment the response team made entry, you were still on the primary access approach."

"Yes."

"That position left you in the wrong quadrant to cover Cors's sector when the response team split toward the eastern corridor."

"Yes," I said. "I should have anticipated the split and positioned for coverage of the eastern approach when Greel gave the push signal. I was oriented on the primary access when the situation changed and it took me four seconds to reorient."

Four seconds. Cors had been in the eastern corridor.

"Four seconds is a significant margin in a split-entry engagement," Vasht said. Not a judgment call. A measurement.

"Yes."

"What would you do differently."

"Position for the secondary cover angle at the push signal rather than maintaining primary focus. The primary access was already covered by two of Greel's people. My position duplicated coverage rather than filling a gap."

"Agreed," Vasht said. "This will be noted as a tactical positioning error and a procedure modification for future engagements." She moved on without lingering, which was correct. The error was named. There was nothing else to do with it except not repeat it. "The Ollen extraction."

"He was hit at the first exchange, eastern corridor edge. I pulled him behind the equipment housing on the western wall, which was the nearest adequate cover."

"The equipment housing was adequate cover, not optimal cover."

"Yes. Optimal cover was the internal cargo framework, fifteen meters further. Ollen couldn't move fifteen meters under his own capacity at that point and I couldn't carry him that distance fast enough to maintain the tactical line. The equipment housing was the right call given the time constraint."

"Sonn concurs with that assessment," Vasht said. "The Ollen extraction is rated as a correct decision under constraint."

I wrote nothing on the mission sheet. I'd already known that. I waited.

"The fire directional call on the gifted soldier's position."

"Yes."

"Walk me through it."

"The gifted soldier was moving along the northern equipment row. I didn't have a direct sighting. The pattern recognition gave me the position signature from the movement pattern at roughly eleven meters. The signature was clean, longer-established integration than mine, running hot from the sprint. I called the position to Greel by equipment row and offset: two rows north, seven meters west of the primary access junction."

"Greel, confirm receipt and action."

"Received the call," Greel said. "Adjusted sighting angle by three meters west from my prior position. Engaged the target at the described coordinates. Single round."

"The call was accurate within two meters of the actual position," Vasht said. "And the engagement was successful because of it."

She let that sit for a moment, not for emphasis, but because she was moving through the data and the data had a shape to it that required that pause.

"The fire directional call represents a use of the Aeluun pattern recognition in active combat conditions at a range and precision level that exceeds Sonn's projected timeline for that capability by approximately six weeks." A brief pause. "I want to be clear about what I'm saying. Six weeks ahead of projection. In live conditions, under fire, with partial view of the target's position."

Greel looked at me from across the room. The look had the quality of assessment, not surprise.

"The conditions were right for it," I said. "The signature was clear. The gifted soldier had been running hard and his integration was running high. That makes the signal denser."

"That's consistent with Sonn's read of the sensor data," Vasht said. "The elevated integration activity created a stronger signature than you would normally encounter at eleven meters."

"Good to know for future reference," I said. "It means gifted soldiers who are pushing their physical capacity are easier to read than ones who are running cold."

There was a short pause on the link. "I'm adding that to the debrief record," Vasht said. "That's a tactical observation that should be in the standing record."

I hadn't thought of it as a tactical observation, just a thing I'd noticed. But that was the difference between what I noticed and what Vasht did with it.

The debrief ran for another forty minutes. Vasht went through the node seizure phase, the data extraction, the extraction sequence, and the secondary corridor approach geometry modification that Denavi had added to the formal procedural record. She asked Denavi specific questions about the terrain analysis, asked Greel about the response team's size and equipment relative to the garrison profile from the pre-mission brief, asked me about the gifted patrol lead's signature characteristics in the initial engagement.

At the end she said: "Operational assessment. The mission achieved its primary objective, the data acquisition, and exceeded the projected intelligence return by identifying the supply routing architecture for four Dominion forward positions rather than the two the brief anticipated. The tactical success is significant. The unit's casualty rate over this engagement, combined with the prior four months of attrition, puts the forward unit below operational threshold for sustained independent operations. I have submitted a replacement rotation request. Estimated arrival, forty-eight to seventy-two hours."

Greel said nothing. Denavi's display was running a quiet update cycle. Ollen had his head back against the wall with his eyes on the ceiling, the way Maret had done two hours before the mission, and I made myself stop that thought before it went anywhere useful.

"Derek," Vasht said. "Off the formal record for a moment."

"Go ahead."

"How many engagements has the forward unit run in the last four months?"

I looked at Greel. He gave me nothing in that look, no signal either way.

"I don't have the full engagement log," I said. "I've been here for one."

"Twenty-three," Vasht said. "Inclusive of last night, twenty-three separate operational contacts."

Twenty-three. I ran the arithmetic without meaning to. The unit had started the four-month period at twelve soldiers and was now at five functional, counting Greel's limited capacity with the arm. Subtract Maret and Cors from last night. Five. Twenty-three engagements over approximately one hundred and twenty days. Two or three contacts a week on average, though the distribution was probably uneven.

The casualty rate wasn't a number I was going to say out loud. Vasht had said it to me because she wanted me to do the arithmetic, not because she expected me to say what the arithmetic showed.

"The data acquisition," I said. "The routing architecture for the four positions. Is Command going to ask for follow-through operations?"

"Within the next forty-eight hours," Vasht said. "Yes."

"With the current unit strength."

"No. The replacement rotation is not optional. But the operational window the routing data creates is measured in days, not weeks. The Dominion will identify the node is dark and begin rerouting resupply within seventy-two hours. The window for exploiting the disruption is the same window as the replacement rotation."

I ran the geometry of that in my head. The replacement rotation arrived at the end of the window. The new soldiers didn't know the terrain. Denavi knew the terrain.

"The two heavier-resourced positions," I said. "The ones on the north approach to the industrial relay network."

"Yes."

"Those are the ones Command is going to want to hit while the resupply disruption is active."

"That is my assessment of Command's likely priority," Vasht said. "It is not a confirmed operational tasking."

"But you're telling me this now so I have forty-eight hours to think about it."

A pause. "The debrief is complete. Formal record closes at 1442. Greel, please ensure the unit's formal casualty reports are filed through the standard channel within the next six hours." Another pause, different texture. "For what it's worth, from someone who has been reading your integration data since the first calibration: last night was not a bad performance for a person who has been doing this for thirty-six hours total. It was a significant positioning error and two correct decisions under pressure. That's a fair balance for a first engagement."

"I know what the balance is," I said.

"Yes," Vasht said. "I know you do."

The link closed at 1442, exactly as noted.

The room was quiet. Greel stood at the display with his right hand on the edge and his left arm in the stabilizer and his face doing nothing in particular. Denavi was looking at the terrain overlay. Ollen had his eyes closed.

Nobody said anything for about two minutes.

Then Durrik's voice came through the secondary comm unit on the equipment case, the monitoring link he'd been running throughout the debrief. He was at the station, not the forward position, and the link quality had the slight processing lag of distance. "Caulfield."

"Here."

A pause. Long enough that I thought the link had dropped.

"Not bad," Durrik said.

The link closed.

I looked at the comm unit for a moment. Then I looked at the mission sheet in my lap, the notation I'd added about the gifted lead's read speed and the secondary corridor coverage.

Not bad. Coming from Durrik, that landed in a specific register that I understood by now. It wasn't encouragement. It was an assessment. A passing grade from someone who didn't give passing grades unless the work had actually passed.

I folded the mission sheet and put it in the kit.

The pattern recognition was running its ambient loop, and the room's six signatures were distributed around me in the usual low-resolution way it tracked things when nothing required active attention. Greel, standing. Denavi, seated at the display. Ollen, seated near the wall. Two perimeter walkers outside the facility's east side, running their circuit at the same interval they'd been running it for four months.

Five signatures where six should have been, and the sixth absence was a thing the pattern recognition registered as a gap the way you register a missing step on a familiar staircase.

I pulled up the notification queue on the unit comm and started writing the formal positioning procedure modification into the standing record. Secondary corridor coverage at push signal. Coverage angle priority over primary approach duplication. It took me twelve minutes to get the language right, which was longer than I'd expected, because the language had to be precise enough to be applied correctly by a soldier I'd never worked with and couldn't brief in person.

When I finished, Denavi looked over at my display without being asked. She read the text, made one change to the third sentence, which improved it, and looked back at her terrain overlay without comment.

I submitted it.

Three hundred forty kilometers away, at the orbital relay position where Vasht's secure link routed through, the debrief data was being processed by Compact Command's operational assessment function. Vasht had already flagged the pattern recognition performance data to Sonn. The flag went in the same packet as the formal debrief record, with two notation tags: "exceeds projected timeline, six weeks" and "active combat verification, high confidence."

I didn't know this at the time. I knew it later, and it wouldn't have changed anything I did between 1442 and the end of the day. The work was the same either way.

Greel turned from the display and picked up his personal kit with his right hand. "Ollen. Medical check at 1600."

"Yes, Commander."

"Denavi. Continue the terrain assessment. I want the north approach position analysis complete before the replacement rotation arrives."

"Understood."

He looked at me. Not a question. Not a task. Just the look of a person who had seen the debrief and the arithmetic and the formal record and was still standing because that was what the next forty-eight hours required.

"The replacement soldiers," I said. "They won't have the corridor context."

"No," Greel said. "They won't."

"Denavi's terrain assessment is the fastest way to transfer the operational knowledge."

"Yes," he said. "I know." He started toward the door, then stopped. "The fire call. At eleven meters, under fire, secondary position."

"Yes."

"I've been in the Caleth Corridor for four months," he said. "That's the first time I've had a fire call that precise from a secondary position with no direct sighting."

He went through the door.

I looked at the equipment case where the mission sheet was folded. Then I looked at the routing display, still running the distribution picture for the four Dominion positions, the geometry of the north approach and the industrial relay network that Denavi had identified as the real objective.

The window was forty-eight to seventy-two hours, and the replacement rotation was forty-eight to seventy-two hours out, and Command was going to ask about follow-through on the routing data within forty-eight hours, and the north approach positions had the highest gifted-soldier pharmaceutical profile in the resupply manifest, which meant they'd be harder to take than the node had been.

Seven soldiers. Of those, five were functional. Of those five, one had a positional error in her record that would modify her approach geometry on all future operations, and that person was me.

Twenty-three engagements. Five soldiers remaining of the original twelve.

I did not say the casualty rate out loud. I'd already done the arithmetic and the arithmetic had finished doing what it was going to do, and now there was terrain to learn and replacement soldiers to prepare briefing materials for and a north approach position analysis that needed to be accurate before the operational window closed.

Denavi pulled up a new overlay on her display without looking at me. "The north approach positions. Their perimeter geometry is documented in the sector survey from six weeks ago. The survey is probably partially outdated."

"Outdated by how much."

"The Dominion has been building in this sector for four months. Six weeks is enough time to modify exterior defensive positions by forty to sixty percent." She ran a comparison overlay, the six-week-old survey against the engagement data from the last several operations. "The exterior approach geometry I can reconstruct from our approach routes and the sensor data from prior engagements. The interior position geometry I can't confirm."

"The routing data might have interior configuration in the supply manifest geometry," I said. "The manifest specified pharmaceutical and nutritional profiles by position. If the resupply routing includes delivery point coordinates."

Denavi pulled up the routing data without being asked, ran a search for delivery point location tags. They were there, three of the four positions tagged with interior delivery coordinates precise enough to indicate internal supply storage locations.

"That's interior configuration data," I said.

"That's interior supply route data," she said. "Close enough." She overlaid the delivery coordinates against the position perimeter geometry from the six-week survey and ran the comparison. The internal routing gave us the position of the supply storage relative to the entry points, which gave us the internal layout's approximate structure. "The north approach primary position has three interior supply routes. That's a minimum of three accessible corridors off the main entry."

"Three corridors is a split-entry problem."

"Yes." She ran the geometry. "Unless you know in advance which corridor connects to the primary command position."

"The heaviest resupply traffic."

She highlighted the delivery coordinate with the highest frequency in the routing data. It sat at the northern end of the interior supply route cluster. "There," she said.

I looked at the geometry. "The north approach primary position's command structure is probably in the northern interior. That's also the most defensible approach from the outside."

"Yes," Denavi said. "That's where they'll put their gifted soldiers." She looked at the display for a moment. "The secondary position has a different internal geometry. Two corridors, lighter resupply. More consistent with a relay position than a primary command post."

"The secondary position is the easier target," I said.

"The secondary position is the one that covers the primary position's western exposure," she said. "You take the primary, the secondary becomes a flanking threat. You take the secondary first, the primary loses its western coverage and you have a clean approach angle." She tapped the overlay. "Four months ago, Greel's unit tried to approach the primary directly. That's where the heaviest losses were."

Four months of prior engagement data in her head, and she was applying it to position geometry reconstructed from supply routing coordinates. I understood why Greel had approved her request to stay past her rotation date.

"Draft this into the terrain assessment," I said. "The approach sequence recommendation, the interior layout inference, all of it. The replacement soldiers need to understand why the sequence matters before they set foot in the sector."

"I'm already drafting it," she said. The same thing she'd said about Vasht's report.

I left her to it and pulled up the secondary task I'd been putting off for three hours: the formal procedure modification to the engagement spread geometry that Vasht had noted as a standing record update. Not the positioning error, which was already filed. The secondary corridor coverage principle as a general operational procedure, applicable across approach-plus-entry engagements regardless of unit composition.

It was paperwork. It was also the thing that might keep the next Cors from dying in the secondary corridor of the next position, because the next forward unit wouldn't make the same mistake if the procedure was in the record.

I wrote it. It took longer than the positioning modification had.

At 1556, Sonn's data flag confirmation came through on the comm. Not a message, just a routing notification that the pattern recognition performance data from the debrief had been received and was being integrated into Derek's development file.

At 1600, Ollen went to medical.

At 1618, a second notification came through. This one was different: a flag from Vasht's relay position, routed through the standard operational channel, marked with the code that indicated Command-level interest. Not a message from Command. A notation that Command had reviewed the debrief and flagged two items for follow-up review.

I knew what the two items were before I opened the flag. The positioning error and the pattern recognition performance. One was a record of a mistake. The other was a record of something that had not been expected.

The flag didn't require a response from me. It was informational. Someone at Command had read the debrief and decided those two items were worth watching.

I filed the flag and went back to the procedure modification.

The routing display ran its quiet loop. The north approach positions sat at their geometry in the sector map, their interior supply routes partially visible through the delivery coordinate data, their gifted-soldier population estimated at four to six each from the resupply pharmaceutical profile.

Denavi worked. I worked. The perimeter walkers outside ran their circuit. The pattern recognition tracked the ambient signatures in the facility and filed them in the low-attention category where it put everything that wasn't moving fast or anomalous.

Somewhere in the secure link infrastructure between the forward position and Vasht's relay station, the debrief data was becoming part of a larger picture that included Sonn's integration readings, Command's operational assessment, and the sensor data from the engagement at the node. That larger picture was being read by more than one set of eyes, and I didn't know yet which eyes those were.

I would find out. Things had a way of becoming clear when they became operationally relevant, and the ones that became relevant most often were the ones I'd already halfway anticipated.

The procedure modification needed two more paragraphs to cover the edge cases. I wrote them. I read them twice, adjusted the language in the second paragraph so it was unambiguous rather than merely clear, and submitted the modification to the standing record.

Then I sat on the equipment case with the display running the routing data and the terrain assessment growing in real time in Denavi's work queue, and I went over the debrief notations one more time, not because I'd missed anything but because the kind of error I'd made at the secondary corridor was the kind that looked like a single incident but had a pattern underneath it, and I needed to know whether the pattern was one I could correct through procedure or one that was going to require a fundamental change in how I read approach geometry under pressure.

Two correct decisions. One significant error. A fire call at eleven meters that nobody expected.

The balance was what the balance was. The next engagement would have different terrain and a different approach geometry and different soldiers in positions I hadn't worked with before.

The only variable I could do anything about between now and then was the procedure. So that was what I worked on.

Outside, the grey industrial sky of the Caleth Corridor ran its afternoon light over the facility's exterior walls, and the perimeter walkers ran their circuit, and forty-eight hours from now there would be new soldiers at the forward position who didn't know the terrain and didn't know the routing data and didn't know that the north approach primary position's gifted soldiers were concentrated in the northern interior near the highest-traffic supply route.

They would need to know all of that before the operational window closed.

The terrain assessment was forty percent complete in Denavi's queue. I pulled up a secondary file and started drafting the integration briefing for the replacement soldiers: what the pattern recognition could do, what it couldn't, how to read a fire call from me when I didn't have a direct sighting, what the call format meant and why the offset coordinates were given in equipment-row units rather than metric distance.

It was the kind of work that didn't feel like war. It was also exactly the work that decided how the next engagement went.

I wrote it carefully, in plain language, the way I'd write anything meant to be understood under stress by someone who'd never seen it before.