Eshet Closes the Distance

The comm call to Vasht took eleven minutes. She listened without interrupting while I laid out the surveillance analysis, then asked three questions: when had I first identified the withdrawal pattern, what was the margin of precision I was assigning to the range calibration, and had I told Greel. I answered all three in order and she said she would flag it to sector command within the hour. No editorial. No escalation in her voice. She filed it the way she filed everything, immediately and without fuss, which was why I kept telling her things.

"One more item," she said, before I could close the link. "The full reading data from Sonn confirms what the preliminary suggested. I am sending the summary to Command with a recommendation that the operational context continue. That recommendation is stronger if the sector activity justifies it."

"You're telling me to stay active."

"I am telling you what my recommendation requires." A pause that was not quite a pause, more a measured beat. "The recommendation was going to Command regardless. This is information."

I closed the link and went to find the training deck.

Durrik ran me through the two-hour rotation without mentioning Vasht or Sonn or integration readings or anything that wasn't the specific tactical problem in front of us, which was how to run a coordinated withdrawal under fire with a unit that had two new soldiers who didn't share its established communication shorthand. The training was physical and specific, and my predictive extension tracked the exercise positions in background mode while I focused on the actual problem, which was Pvt. Hetch moving at eighty percent of his stated capability and covering it well enough that Durrik hadn't noticed but I had, because the pattern recognition read stride intervals even when I wasn't asking it to.

I mentioned this to Durrik after the rotation. He noted it without expression and scheduled a physical assessment for Hetch the following morning.

The next two days were quiet. Not operationally quiet, the unit ran a standard patrol pattern and Greel kept the forward position active, but quiet in the sense of no unexpected contact. The Dominion's presence in the sector was a background frequency I'd been tracking for weeks, a rhythm of patrol intervals and supply movement that varied within a predictable range. On the morning of the third day, the rhythm changed.

I was at the forward position's secondary vantage point when the shift registered, the pattern recognition flagging a density increase in Dominion movement signatures to the northeast that didn't fit the established interval. Not a patrol. Too many signatures, too close, with the compressed spacing of units moving in coordination rather than the loose spread of routine activity.

Greel was twenty meters away, reviewing a terrain schematic with Denavi, when I got to him.

"Northeast sector," I said. "Multiple units. Coordinated spacing."

Greel looked up from the schematic. He had the particular stillness of a commander who doesn't show reaction times and was reading terrain data in about three seconds instead of the fifteen it would take to process a full schematic.

"Numbers?"

"Pattern suggests four to six distinct unit signatures. At least two with the gifted-soldier movement profile." I thought through what the spacing meant. "They're not in approach formation. They're settling into position."

Denavi straightened from the schematic and looked northeast automatically, which was the wrong direction from this interior vantage point but was the reflex of four months in this sector. "Setting up observation?"

"Or staging."

Greel folded the schematic and handed it to Denavi. "Get the unit up. Quiet." He turned to me. "How long have we had?"

"The shift registered eight minutes ago. The movement before that, settling-in activity, that's been building since before dawn. I was reading it as normal variation until the density crossed a threshold."

Greel's expression didn't change but he looked at me for one second longer than the information required, which I'd learned to read as a form of question. "You're saying this has been building since before dawn and the threshold registered eight minutes ago."

"Yes."

"What does the building pattern look like?"

"Slow. Deliberate. Not the spacing you'd use if you were worried about being detected." I organized the data from the last several hours. "Whoever is directing this isn't concerned about our sensor range. They know what we're running and they've positioned accordingly."

Greel moved. He was already sending the quiet-alert signal to the unit before his next step landed. "Get to Vasht. I want sector command notified."

I went.

The Dominion engagement came three hours after Greel's alert, and it came from the south, which was not where the northeastern staging had suggested. The probing force was small: two gifted soldiers and four standard infantry, moving against the forward position's southern approach in a textbook two-angle press that Greel's repositioned unit absorbed without major difficulty. The fight was sharp and brief. One Compact soldier took a grazing hit on his left arm, treated in the field in ten minutes. The Dominion force withdrew after four minutes of contact, taking their casualties with them, which was operationally disciplined and told me something.

Denavi, running the post-contact perimeter check, came back with a sensor log from the southern approach. "They left a passive package," she said. "Behind the second overhang. Small one." She held up a component the size of a coin with a flat antenna array along one edge.

Greel examined it. "Surveillance."

"But they know we found it," I said.

"Do they?"

"The withdrawal was clean. They covered their exit and they left a package they knew we'd sweep for." I looked at the sensor log in Denavi's hand. "They didn't leave it because they thought we'd miss it. They left it to see how we handle it. Whether we move it, destroy it, or leave it in place tells them something about our protocol."

Greel held the device for another moment. "Leave it in place. Move nothing." He handed it back to Denavi. "Put it back exactly where it was."

Denavi went. I watched the northern sector while she did it.

The second probing engagement came six hours later, at dusk, from the northwest. Larger this time: four gifted soldiers, seven standard infantry. The attack pressed two positions simultaneously and held contact for nine minutes before withdrawing. Maret's replacement, a Chethari soldier named Bael who had been with the unit for three weeks and was steady in the way of someone who'd seen enough not to be surprised by much, took a hit in the second minute that knocked him off one position. I redirected coverage automatically without radio call, reading the gap before Greel could respond to it.

The gap held. Bael got up on his own, which told me the hit was armor-distributed rather than penetrating, and retook his position before the contact ended.

The Dominion withdrew again. Orderly, covered, no abandoned casualties.

Greel's debrief with the unit was short. He covered the engagement facts, noted Bael's recovery, and gave the next watch assignments. When the unit dispersed he looked at me across the forward position's interior and waited.

"Two probes in one day," I said. "Different axes. Different force compositions. Both withdrew on a clean signal, not under pressure." I had been organizing the pattern since the second contact ended. "They're not testing our defensive capability. They know our defensive capability from the previous engagements and whatever sensor data they've been pulling on my signature."

"What are they testing?"

"How we respond to information uncertainty. Whether we reinforce south after the first contact, which would tell them something about our read on their positioning. Whether we adjust the sensor protocol when we find the passive package. How fast we respond to a northwest axis versus a south axis, which gives them our internal communication timing." I thought through it. "They're building a decision-tree profile on the unit's command responses."

Greel was quiet for a moment. "The surveillance analysis you brought me from the equipment bay. Eshet."

"The quality of this operational design is consistent with the surveillance precision. Same methodology." I looked at the sector map on Greel's workstation. "Eleven years of integration gives you a read on how enhanced pattern recognition works from the inside. She knows what data my system collects and what it doesn't. She's designing the probes to generate data she can read while staying in the gaps I can't read well."

"What are your gaps?"

"At extended range, my pattern recognition gives me movement and spacing data but not intent. I can tell you how many, where, and how they're moving. I can't tell you the mission objective until the movement pattern resolves. She knows that. The probes today were at extended range." I put my finger on the map at the northeast staging area. "She's got assets here that I've been reading for twelve hours. They haven't moved. That's not a staging force. That's a frame."

"A frame for what?"

"For the probes. If I were trying to run this, I'd use a large stationary presence to hold my target's attention at one vector while I ran the actual information-gathering operations from a different direction. I keep looking northeast because there are signatures there. The probes are coming from south and northwest." I took my finger off the map. "She's been standing in the northeast with four to six units doing nothing for twelve hours so I will keep reporting northeast as the threat, and the actual data-collection runs come from the directions I'm not fixated on."

Greel studied the map. He had the face of someone recalculating several things simultaneously. "You're fixated on northeast."

"I was. I'm not now." I traced the pattern of the two probes on the map. "South and northwest. If she runs a third probe, I'd expect east. Complete the perimeter survey. One more data point and she'll have a full positional profile."

Greel looked at the map for another ten seconds. He put his hand on the communication panel. "I'm going to request sector command reinforce our east-facing coverage before tomorrow dawn."

"Tell them the reinforcement request is based on an anticipated third probe, not a confirmed contact. If command thinks we're calling in a crisis, they'll want to pull us back."

Greel sent the request with that framing. I went back to the secondary vantage point and ran the northeast signature through the pattern recognition while the ear tone ran its frequency and the predictive extension sat in standby, waiting for something close enough to work on.

The northeast signatures weren't moving. Hadn't moved in thirteen hours.

She was patient. Eleven years in, running a ground campaign in a contested sector against a Compact unit she'd been grinding down for four months, and she was patient enough to stand four to six units in a static position for a full day just to establish a false priority axis. I filed this. Whatever the probes were collecting, patience was part of the profile she was building on us, and whoever had stood four to six units still for thirteen hours wasn't going to make the final engagement on a schedule driven by anything except her own read of the data.

I tracked the northeast position until 2300 and then handed the watch to Bael and went to get three hours of sleep. The predictive extension ran standby tracking through all of it, which I was accustomed to now. The sleeping brain processes things in the background without full consciousness, and the extension had fitted itself into that processing space the way a tool fits into the hand that uses it regularly.

The third probe did not come from the east.

It came at 0340 from the north, which was the one axis neither Greel nor I had predicted, and it came with eight gifted soldiers moving in a coordinated spread pattern that my pattern recognition read two and a half seconds before the unit's sensor equipment registered the contact. I was awake and moving toward Greel's position before the sensor alert sounded.

The engagement lasted fourteen minutes. The unit held all positions. Two Compact soldiers took hits, both survivable. The Dominion force withdrew after fourteen minutes on a clean signal.

And they left no passive package this time.

Greel and I stood in the forward position's interior at 0420 while the unit ran the post-contact protocol, and Greel said: "North."

"She had four probes available and she ran three." I thought through the geometry. "South, northwest, north. She skipped east. Either she doesn't need the east data, or she already has it from a source I didn't account for."

"The passive package," Greel said.

"She left it facing south but the array isn't directional if the processor is good enough." I ran back through the assumption. "We read it as south-facing because of the antenna orientation. If she's running a better processor than standard sector-grade, she's getting omnidirectional data from a device I read as directional."

Greel looked at the southern approach.

"She has the east profile," I said. "Has had it since the first probe. The north engagement tonight wasn't a data run. She was checking whether her east data was accurate by running a north probe and seeing if we'd reinforced east based on my prediction." I worked through the logic. "We didn't reinforce east. We reinforced east-facing coverage, which is different. Our sensor posture changed without our physical position changing. She read the sensor posture change in the engagement pattern and got a data point on how my predictions affect the unit's responses."

Greel was quiet for six seconds. "She's not just building a profile on the unit. She's building a profile on you."

"On how the unit responds to my input." I looked at the map. "Which is a profile on me. She wants to know how my predictions translate into command decisions, so she can calibrate the gap between what I can read and what it takes to get a unit acting on that read."

The prediction accuracy, the command communication timing, the sensor posture response, the passive package protocol. She had data on all of it now.

Greel sent a full contact report to sector command at 0430. I wrote a supplementary assessment that laid out the profiling pattern and attached it to his report. I didn't soften the conclusion: whoever was directing the Dominion operations in this sector had been running a systematic intelligence collection program for at minimum seventy-two hours, and the methodology was consistent with an operator who had significant personal experience with Aeluun integration and was using that experience to design data-collection operations specifically calibrated to the gaps in my pattern recognition coverage.

Vasht reached me by link twenty minutes after the report went up.

"The supplementary assessment," she said. "You're confident in the attribution?"

"Confident in the methodology signature. The attribution to Eshet is inference."

"But your inference is."

"Based on eleven years of integration and the operational quality. Yes." I let that stand on its own. "Sonn described Eshet's integration level in the briefing materials. Someone running that deep has enough familiarity with how the system works to design around it. The precision of the coverage gaps in these probes isn't lucky placement. It's calculated."

Vasht was quiet for a moment. "Compact Command has signals intelligence suggesting Eshet relocated from her previous sector command position forty-eight hours ago. No confirmed destination." Another short pause. "This is consistent with the Caleth Corridor."

The northeast staging position. Sitting still for thirteen-plus hours because she could afford to be still and because her patience was part of the data she was generating, showing me a large static presence I would keep checking so the probes could run where I wasn't looking.

"She's here," I said. "On the ground. The probes have a tactical signature that's not consistent with remote command."

"I am flagging this to sector command and updating Command's assessment of the Corridor threat level." Vasht's voice had the quality of controlled urgency she used when something had moved from manageable to actively pressing. "The updated threat level will likely trigger a reinforcement request. That request takes time."

"How much time?"

"At current allocation priority, sector reinforcement for the Corridor is a seventy-two-hour response minimum." She didn't editorialize. "This is the information you have."

Seventy-two hours. I looked at the map on Greel's workstation.

"Derek." Vasht waited until I looked at the link's camera. "Eshet has been building a capability profile on you for weeks. What she has now, after three probing engagements in one day, is a solid read on your pattern recognition and how your command communication integrates with the unit's responses. What she does not have confirmed data on, based on what you've told me, is the predictive extension."

She was right. The probes had been calibrated precisely to my documented range and capability. Which was the capability profile built from the signature data generated before the predictive extension went structural.

"The modification," I said. "The signature change from the range extension."

"Sonn says the predictive extension generates a distinct signature overlay on top of the modification's baseline. If Eshet's intelligence is working from pre-extension signature data, she has an accurate picture of what you were three weeks ago." Vasht let this settle. "She does not know what you are now."

I thought about the probes. The precision of the coverage gaps, designed to stay outside pattern recognition range. The passive package left with a processor good enough to read omnidirectional data, feeding her a continuous stream of sensor and positional information.

She had built a detailed model of an asset that no longer exactly existed.

"She's going to engage on that model," I said.

"She will design her main operation around the gaps she has catalogued. Yes." Vasht's expression through the link was controlled and specific. "I am not telling you what to do with this information. I am making sure you have it."

"I know."

She closed the link, and I went to find Greel.

He was at the perimeter, doing his own assessment of the north approach, when I reached him. The sky was moving toward the gray that precedes dawn by forty minutes, and the northeast sector had been static for sixteen hours now, the signatures sitting in their positions like they had nowhere better to be.

"She's on the ground," I said. "Vasht has signals intelligence consistent with a relocation to this sector forty-eight hours ago."

Greel turned from the perimeter. "Confirmed?"

"No confirmed destination, but the operational methodology of the last twenty-four hours is consistent with personal command." I watched the northeast sector. "Reinforcement request is minimum seventy-two hours from confirmation."

Greel was quiet for a moment. "What does she have?"

"A detailed profile on my pattern recognition range and coverage. My command communication timing. How the unit responds to my input under pressure." I paused. "She does not have confirmed data on the predictive extension. The signature data she's been reading predates it."

Greel processed this the way he processed tactical information, quickly and without asking for it to be simpler. "She'll design the main engagement around the gaps she catalogued."

"Yes."

"Which are no longer your actual gaps."

"Correct."

Greel looked at the northeast sector for a long moment. "How long before she has enough data to run the main engagement?"

I thought about the probes, what she'd collected, what was left to collect. Three probes in one day, each one systematic. She wasn't rushing. She wasn't stalled.

"She has what she needs," I said. "The fourth probe didn't come. The north engagement tonight wasn't a data run, it was a confirmation check. She's done collecting." I looked at the northeast staging signatures, sitting still in the early-morning gray, patient and unmoving. "She's ready when she decides she's ready."

Greel turned and walked toward the forward position's interior. "Get some sleep. Three hours. I'll post the watch."

I looked at the northeast sector one more time. Sixteen hours without movement, the signatures sitting in their positions with the quality of something that was done waiting.

The ear tone ran its frequency in my right ear.

Three hours of sleep, and Eshet was ready, and seventy-two hours was a long time.

I went inside.