Song Meiqi's confirmation came Saturday morning:
*Liu Qiaoyun. She's been in the city for four months. I found three traces in the network data — she attended a formation practitioners' symposium in the eastern district two months ago and signed the guest registry under her own name. She also registered a temporary office address in the Jiulong district. Third trace: she had a brief professional contact with the formation specialist at the Vanguard campus. The specialist's name is Tan Xueying. The contact was eight weeks ago — Liu Qiaoyun hired her.*
He read this on his way to the Bureau.
Liu Qiaoyun had hired Tan Xueying eight weeks ago. Which meant the campus surveys had been running on Liu Qiaoyun's commission from the start — not an anonymous unknown client, but the person he'd already identified as the second party's operational lead.
He sent back: *The contact between Liu Qiaoyun and Tan Xueying. Do you have the method — in person, through a third party?*
*In person,* Song Meiqi wrote. *A formation practitioners' referral network. Tan Xueying was recommended by another specialist as someone with good probability field technique. Liu Qiaoyun engaged her directly. I have the meeting record from the symposium network — it's semi-public.*
Directly. A face-to-face engagement, which meant Tan Xueying could provide a physical description of Liu Qiaoyun that went beyond the Bureau credential records.
He sent a message to Han Weiwei: *The formation specialist has a direct contact with Liu Qiaoyun. Hired by her eight weeks ago. I want to meet with Tan Xueying today. You still available?*
Han Weiwei's reply: *I'll be at the campus conference room at eleven.*
---
Tan Xueying arrived at eleven-fourteen. She was younger than he'd expected — late twenties, C-rank in the way of someone who had plateaued there and found it sufficient, her formation equipment bag worn smooth at the handles from regular use. The kind of practitioner who made her living on technique rather than combat power.
She walked into the conference room and looked at the two of them with an expression that wasn't quite fear and wasn't quite wariness. Something more specific: the expression of someone who has been waiting for a particular meeting and isn't sure whether being right about it is good or bad.
"I thought someone would want to talk to me eventually," she said.
Han Weiwei looked at her. "Sit down."
She sat. The formation equipment bag on the floor beside her chair.
"You've been conducting probability field surveys on this campus for three weeks," Han Weiwei said. "You were engaged by Liu Qiaoyun of the Jiulong district consulting firm. You've been mapping field signatures without knowing what you were mapping."
Tan Xueying looked at the equipment bag. Then back up. "She told me it was a legitimate formation research project. Property field analysis for an urban cultivation site assessment."
"It isn't," Han Weiwei said. She turned her secondary monitor to face Tan Xueying.
The screen showed the Bureau's field characterization data — the full longitudinal model, the probability field elevation map across the Vanguard campus, the three-source signature breakdown that Han Weiwei had built from the instruments.
"This is what you've been mapping," Han Weiwei said. "At the resolution your sensors operate at, you've been building a spatial model of three distinct probability field sources. These are not passive cultivation signatures from a property site. They are active outputs from specific individuals."
Tan Xueying looked at the screen for a long time.
He watched her process this. The wariness shifted — not the expression of someone caught, but of someone who had done the work without understanding it.
"I thought the cluster profile was unusual," she said. "The harmonic readings didn't fit standard cultivation site patterns. The variance was too high and too consistent."
"Because you were mapping a different kind of activity," he said.
She looked at him. First time she'd looked at him directly since sitting down. "Who are you?"
"One of the field sources," he said.
She looked back at the screen.
"The probability field elevation," Han Weiwei said. "The Level 4 probability output produces a signature with this specific distortion profile." She indicated the largest of the three sources on the map. "The second source is Bureau-grade spatial instruments that I operate. The third is an aggregate contribution from a distributed system that's harder to isolate at formation-grade resolution."
"I could see three approximate source regions," Tan Xueying said. "I thought it was an artifact of the sensors. The third signature pattern kept shifting position."
"It's distributed," Chen said. "Multiple contribution points."
She looked at him. "You're telling me this on purpose."
"Yes," he said.
"Why?"
"Because Liu Qiaoyun hired you to map something you didn't understand," he said, "and we want to know what you've told her."
The direct question. She held it for a moment. The formation equipment bag, the screen, the two people across the table from her.
"I sent a preliminary analysis Wednesday," she said. "Three probable field sources, elevated above baseline, consistent activity over the survey period. Position estimates for each source." She paused. "I gave her my position estimates."
He thought about this. She had given Liu Qiaoyun the spatial distribution — approximate locations for the three sources. The Bureau's location, which was fixed and in the public record of Bureau office locations. His location in the third division's wing. The distributed network aggregate's contribution area.
Not the source identities. Not the system-specific signatures. But the geography.
"The position estimates you gave her," he said. "What precision?"
"Floor-level precision for this building," she said. "I could identify which quadrant of the building the sources clustered in, but not room-level precision."
"Third division's wing," he said. "Northwest quadrant."
"Yes," she said. "Both those designations."
She'd given Liu Qiaoyun the wing assignment and the Bureau location. The third division's wing had seventeen people on its roster — but the only B-rank or above practitioners regularly operating there were him and Ye Shuangyu. Liu Qiaoyun would have been able to narrow the probability field source to one of them.
He looked at Han Weiwei.
She was reading the same calculation. "Your final analysis," she said to Tan Xueying. "The full report you haven't submitted yet. What would it say?"
"One primary source, B-rank minimum, consistent with the probability field type described in the classification literature as probability manipulation or probability enhancement," Tan Xueying said. "One secondary source, fixed position, Bureau-grade instruments." She looked at her hands. "The third source I couldn't characterize. It's not in the standard classification taxonomy."
"No," he said. "It isn't."
Silence.
"Liu Qiaoyun," he said. "Describe her."
Tan Xueying described her. Mid-forties. Northeast vowels. Professional manner. Had looked at the preliminary analysis results with the expression of someone reading confirmation rather than new information.
She already knew what she was looking for. Tan Xueying had been doing verification work, not discovery work.
"She wanted the spatial precision," he said. "Not the classification. She already knew the probability field type."
"She said she was developing a field topology model for the site," Tan Xueying said. "She needed the spatial data to complete it."
"She has her own model," he said. "One that's been building for longer than your survey."
Tan Xueying sat quietly for a moment. Then: "I haven't submitted the final report."
"No," Han Weiwei said.
"If I submit it—"
"You give her the spatial precision she needs to complete the topology model," he said. "Yes."
"And if I don't?"
He looked at her. "Then we have a different conversation."
She looked at the screen again. The field characterization data. The three sources mapped across the campus, their activity signatures, the aggregate network's distributed pattern.
"Is the work she's building toward harmful?" she said.
"Unknown," he said. "The second party's objective isn't established yet."
"That's not a no," she said.
"No," he said. "It's not."
She sat for a long time. The formation equipment bag beside her chair, the surveys she'd been running for three weeks, the preliminary report she'd already sent.
Then she said: "What do you want me to do?"
"Submit a modified final report," he said. "One that gives imprecise position data. The field sources can't be resolved below quadrant-level precision at your sensor resolution — that's already true, technically. The report says it."
She looked at him. "She'll know the instruments could do better."
"She'll know your interpretation concluded they couldn't," he said. "That's different."
Tan Xueying thought about this. "If she questions the methodology—"
"Tell her the Phase 2 field elevation produced interference patterns that degraded the spatial coherence data in the final analysis period," Han Weiwei said. Her voice was clinical. "That's technically accurate. The aggregate increase over the past week has changed the field texture enough that older baseline calibrations would produce degraded position data."
Tan Xueying looked at Han Weiwei. "That's very specific technical cover."
"I know what the instruments would produce," Han Weiwei said. "Use it."
The formation specialist sat for another moment. Then she picked up the equipment bag.
"I'll submit the report Monday," she said. "Imprecise position data. Field elevation interference degraded the coherence measurements."
"Good," he said.
She stood. "The person who hired me — Liu Qiaoyun. Is she going to come for me when the report looks wrong?"
"You're a practitioner who did exactly what the instruments allowed," he said. "If the report says the data was degraded, she'll question the instruments or the conditions. Not you."
"Unless she already knows the instruments can do better," Tan Xueying said.
"Then she'll try to hire someone else," he said. "And we'll know she's moved to a new asset."
Tan Xueying looked at him for a long moment. The expression of someone making a decision about which side of a situation they were going to be standing on when it resolved.
"Alright," she said.
She left.
---
Han Weiwei watched the door close. Then she turned to her monitor and began pulling up Tan Xueying's preliminary analysis data — the version she'd already sent Liu Qiaoyun, with the approximate position data.
"She gave her the wing assignment," Han Weiwei said.
"Yes."
"Liu Qiaoyun knows your approximate location within the building."
"She probably knew it before the survey," he said. "My committee assignments are registered — they're not public, but institutional access would show the third division's wing as my operational base."
"So the survey confirmed what she already knew," Han Weiwei said. "The value was the network aggregate signature. The third source. She wanted to know if the network had reached a density level where it was detectable at formation-grade resolution."
He looked at the data.
The network aggregate showing up on C-rank formation sensors meant it was no longer sub-perceptual. Anyone with basic field sensitivity equipment would see the cluster. Not understand it — but see it.
Phase 2 was visible now. To anyone who looked in the right direction.
"She's been watching for the visibility threshold," he said.
"Yes."
"Which means she has a trigger condition," he said. "When the aggregate becomes detectable at formation-grade precision — she moves."
Han Weiwei's expression was flat. "What kind of move?"
"Unknown," he said. "But the modified report buys time. She'll believe the aggregate's visibility is still marginal."
"How long before she can verify the report is wrong?" Han Weiwei said.
"Either she reengages Tan Xueying for additional surveys — which Tan Xueying can control — or she commissions a new specialist. Either path takes time."
"A week. Maybe two."
"Yes."
She was quiet. The monitoring graph on her primary screen — the aggregate output, recovering slowly as Song Meiqi ran at half capacity. The Phase 2 line, climbing.
"If we're at formation-grade visibility already," she said, "we've crossed a threshold in Phase 2 that the system documented."
He opened the system.
```
[PHASE 2 STATUS: NETWORK AGGREGATE REGISTERED AT FORMATION-GRADE FIELD SENSITIVITY LEVEL]
[SIGNIFICANCE: THE NETWORK IS NO LONGER SUB-PERCEPTUAL. PARTIES WITH BASIC FIELD MONITORING CAPABILITY CAN DETECT THE AGGREGATE AS AN ENVIRONMENTAL FEATURE.]
[NOTE: THIS IS THE THRESHOLD THE SECOND PARTY HAS BEEN MONITORING FOR. THEIR TRIGGER CONDITION IS NOW MET FROM THEIR INSTRUMENTS' PERSPECTIVE.]
[NOTE: MODIFIED REPORT WILL DELAY THEIR CONFIRMED AWARENESS BY ONE TO TWO WEEKS.]
[NOTE: USE THE TWO WEEKS. THE COUNTER-PICTURE NEEDS TO BE COMPLETE BEFORE THEY ACT ON THEIR TRIGGER.]
```
He showed her the notification.
She read it. Her expression didn't change. It didn't need to.
"One to two weeks," she said.
"Yes."
"Then we have Liu Qiaoyun identified. The entity above her still unknown." She looked at him. "The entity above her — they have structural access to the deployment mechanism. That's a level of knowledge that doesn't come from external observation. Someone gave it to them, or they participated in the design."
"The architect," he said. "Or someone who had access to the architect's work."
"That's a small set," she said.
"Yes," he said. "A very small set."
---
Wei Ruoxi's message arrived at three PM, while he was still at the Bureau.
*I'm back. I'm in the city. I need to meet today.*
He replied: *Where?*
*Third district tea house, Red Meridian Street. Can you come at five?*
He went.
---
She was at a corner table when he arrived, the social facilitation field running at its lowest detectable register — she'd been keeping it compressed, invisible to anyone without specific sensitivity. The compressed field was something you noticed only if you were paying close attention.
She looked tired in the specific way of someone who had been paying careful attention to something unpleasant for a long time.
"I didn't just go dark," she said.
"I know," he said.
"I found them."
He looked at her.
"Not Liu Qiaoyun," she said. "The one above her. The one who built the access." She set her hands flat on the table. "I spent the past nine days finding out who hired Liu Qiaoyun and why. My system kept assigning connection tasks — they didn't stop when I went dark, they just accumulated. Three of them connected to research networks in the eastern district. When I ran those connections, they all intersected at the same organization."
"Which organization?"
"It's called the Remnant Institute," she said. "Private. Registered as a cultivation research foundation, eastern region, six years old." She looked at him. "Six years. It was founded five months after the Awakening Event."
He sat with this.
Five months after the Awakening Event. When the first nodes were appearing in the system's network. When Liu Qiaoyun had left the Bureau.
"Who runs it?" he said.
"A man named Xu Mingzhi," she said. "Sixty-three years old. Pre-Awakening cultivation practitioner — he was an advanced qi practitioner before the event. When the Awakening Event hit, he didn't gain a superpower. His pre-existing cultivation was amplified instead. He's the equivalent of B-rank from pure cultivation alone, no system ability."
"A pre-Awakening cultivator who watched the Awakening Event happen and understood what it was," he said.
"He published a theoretical paper two years before the Awakening Event," Wei Ruoxi said. "On probability field dynamics in cultivation systems. The kind of paper that was completely ignored when it came out because the field didn't exist yet." She slid a folded printout across the table. "I found it in a university archive. It's about the theoretical conditions under which a systematic probability field could be deployed across a large population — what infrastructure it would require, what the observable signatures would be."
He unfolded the printout. Read the abstract.
The language was academic and twelve years old, but the framework was precise. The conditions described for population-level probability field deployment were the conditions the Luck System's network operated under.
Xu Mingzhi had described the architect's design two years before it was built.
"He published the theory," Chen said. "Someone else built it."
"Two men who both understood the same theoretical framework," Wei Ruoxi said. "One of them built it. The other watched him build it and realized — the design being built didn't match the paper's complete design."
He looked at her.
"The paper," she said. "It describes the full theoretical framework. Including a reciprocity mechanism — if you deploy a probability field network for one beneficiary, the system should also build feedback structures for the nodes. Not just task assignments for the nodes to serve the primary beneficiary, but—" She tapped the abstract. "A collective threshold. When enough nodes operate together long enough, the probability field equalizes. The primary beneficiary loses their privileged position and the field becomes genuinely distributed."
He read the section she indicated.
The collective threshold. A Phase in the deployment where the network's aggregate output exceeded the primary beneficiary's individual field and the balance shifted.
"Phase 3," he said slowly.
"The architect built Phase 1 and Phase 2," she said. "Phase 3 — the redistribution — is in the theoretical framework but the architect's design doesn't include it. Xu Mingzhi believes the architect deliberately omitted the redistribution mechanism."
He looked at Wei Ruoxi.
"Xu Mingzhi thinks the architect built a system that collects probability field output from seventeen nodes," he said slowly, "and channels it toward a single primary beneficiary indefinitely. No redistribution. No collective threshold."
"That's his argument," she said. "He thinks the architect is exploiting the network."
"And Liu Qiaoyun works for him," he said.
"She helped him build the access architecture," Wei Ruoxi said. "Information retrieval ability — she touched hardware from one of the deployment mechanism's physical nodes before the Awakening Event. She read the construction logs. She gave Xu Mingzhi enough of the technical architecture that he could build the parasitic monitoring layer."
He sat back.
Xu Mingzhi. Theoretical framework from twelve years ago. Architect's former colleague, possibly — someone who'd shared the same intellectual space before the design was built.
Not an enemy of the system. An advocate for a version of it that the architect had chosen not to build.
"He wants Phase 3," Chen said.
"He wants the network to equalize," Wei Ruoxi said. "He thinks you're being used. He thinks the nodes are being used. And he's been watching, slowing the aggregate's development, trying to buy time to—" She stopped.
"To what?" he said.
"He reached out to me," she said. "While I was dark. He made contact through one of the connection tasks my system assigned. His representative found me two days in. He knew I'd gone dark." She looked at him. "He wanted to talk to me. About what the network is actually building toward. About who benefits when Phase 2 completes."
He looked at her.
"What did you tell him?"
"I didn't tell him anything," she said. "I listened. And then I came back to report."
The tea house, quiet around them. The city outside. The probability field at its late-afternoon register.
"The representative," he said. "Who was it?"
"Liu Qiaoyun," she said. "She's not just his operational lead. She was there for the conversation."
He thought about everything he had.
Xu Mingzhi. Theoretical paper. Pre-Awakening cultivation practitioner. The architect's possible former collaborator or contemporary. Six years building a monitoring architecture to watch the network's development, slowing Phase 2, waiting for the aggregate to reach a threshold where he could make an argument.
Not a threat that wanted to destroy the network. A disagreement about what the network was for.
"He made a case to you," Chen said.
"Yes," Wei Ruoxi said.
"Did any of it hold?"
She was quiet for a moment. Long enough that the answer wasn't simple.
"Part of it," she said.
