Rowan's file on Yoon Kang ran forty-seven pages by the end of the first week of recovery.
He read it twice — once for content, once for what the content implied. The shape was different from what he'd known in the original timeline in ways that weren't just about timing. Yoon Kang in the original timeline at year three had been a mid-tier operation: predatory lending, awakener debt collection with muscle, occasional labor coercion for lower-ranked hunters who owed more than they could repay through cash. Unpleasant, not particularly sophisticated, manageable through standard channels if you had the right Association contacts.
Yoon Kang in this timeline at the one-year mark was smaller in visible footprint and larger in actual reach. Less muscle, more infrastructure. The lending operation existed, but the file made clear it was partially a front for something Rowan had only partially mapped: access brokerage. Yoon Kang's network controlled access to certain D-rank and E-rank dungeons in the eastern and northern districts through a system of unofficial "clearing fees" paid by smaller parties who couldn't afford disputes. The awakener black market was layered — Nadia Voss controlled the premium tier, several mid-tier operations competed for everything below that. Yoon Kang had carved out a niche that was less about goods and more about permission.
Which was a more sophisticated model than the original timeline version had run. And more difficult to approach directly.
The Blackwood family debt was in the lending arm. Rowan had found the connection through the supplementary documentation from the debt-assistance application. The obligation was three years old — it predated the Awakening by eight months, which meant the Blackwood family had borrowed from Yoon Kang's precursor operation before the world changed. When the Awakening hit and the economy collapsed into chaos, the legitimate debt assistance programs had caught some of the ordinary commercial debt. The Yoon Kang portion had survived the chaos without resolution because it was never in the formal system.
Current amount: approximately one hundred and forty thousand, with compounding interest terms that Rowan described, in his typically flat analytical tone, as "predatory in the technical sense of the word." The principal was from three years ago. The current amount had grown from an original loan of around fifty thousand. Yoon Kang's repayment terms were structured specifically to ensure the debt grew faster than it could be repaid.
Standard practice for the operation.
The Blackwood family was not behind on payments. That was the second piece. They were current — Sera's father was making regular, painful payments that were keeping the loan from triggering the collection clause. The collection clause, which Rowan had obtained through an indirect source in the lending network, required Yoon Kang's people to "secure alternative value" if cash payments fell below a certain threshold for two consecutive cycles.
Alternative value. That phrase.
He called Rowan.
"The alternative value clause," he said, when Rowan answered. "Has it been invoked."
"Not formally. But there have been two visits to the Blackwood household in the past six weeks." Rowan's voice had the quality of someone presenting bad news carefully. "The first visit was standard collection contact. The second visit, three weeks ago — the representatives asked about Sera specifically. Her class potential. Whether she'd awakened yet."
"They're assessing her."
"The alternative value framework in Yoon Kang's contracts covers 'services rendered by affiliated awakeners.' It's a labor coercion clause. If the debt can't be paid in cash, the debtor can provide awakened labor services to the network's operations." A pause. "Sera hasn't awakened yet. But she will. And when she does, if the cash payments are still barely meeting the threshold, they'll make the offer."
"It won't be framed as coercion."
"It'll be framed as an employment opportunity that happens to also service the debt obligation." Another pause. "That's the model. The debtor family feels like they're choosing. They're not choosing."
He sat with this.
In the original timeline, Dorian had coerced Sera through the debt — but through a simpler version of the same structure. Leverage, threat, compliance. The mechanism here was more elaborate: Yoon Kang built debt traps that produced compliant labor on a schedule. By the time Sera awakened with the Iron Fortress class, she'd have been managed into the situation long enough that refusing would feel impossible.
And Dorian, in this timeline, was already affiliated with Fenner's network. Which was connected to Crane's infrastructure. The Association's compliance division.
The two coercion vectors were separate but not unrelated. Yoon Kang provided the structural pressure. Crane's network provided the institutional access. If those two connected through Dorian—
"Rowan. Has there been any documented contact between Yoon Kang's operation and Crane's compliance infrastructure."
A pause. "I don't have that in the current file. But I haven't specifically looked for it." Another pause, longer. "If you're suggesting that Yoon Kang is functioning as a talent identification and coercion pipeline for Crane's network—"
"I'm suggesting it's worth mapping."
"I'll map it." He could hear Rowan thinking. "This changes the ethics review documentation scope significantly if there's a connection. We'd be looking at an organized system rather than a compliance officer with a conflict of interest."
"We were always looking at an organized system. We just didn't have the full shape."
"Right." A pause. "The Blackwood family. If we want to clear the Yoon Kang debt before the alternative value clause becomes active, we need to move before Sera awakens. Her awakening is—"
"About five months away."
"That's the original timeline estimate."
"Yes." He looked at Rowan's file. "In this timeline, with the changed conditions — earlier, probably. The Awakening process responds to stress and environmental factors. The Blackwood family is under significant financial stress."
"Which could accelerate the class manifestation."
"Which could accelerate it. So we have fewer than five months and possibly considerably less." He closed the file. "The debt is one hundred and forty thousand with a predatory structure. That's not a number I can route through standard channels without it looking like someone with resources stepped in. And if Yoon Kang sees an external party clearing their debt obligations, they'll know someone is working against their coercion pipeline."
"Which triggers them."
"Which triggers them." He looked at the ceiling. "So I can't just pay it off. I need to pay it off in a way that doesn't look like I'm paying it off."
A pause.
"That's a complex financial operation," Rowan said.
"It is." He paused. "Nadia Voss."
The name sat in the air for a moment.
"The black market queen. She and Yoon Kang—" Rowan stopped. "She controls the premium tier. They're in competition for parts of the mid-tier. She'd have reason to want a Yoon Kang debt obligation disrupted."
"Or she'd have a framework for routing a debt clearance in a way that looked like a legitimate transaction between operations rather than an external intervention." He thought about Nadia Voss. He'd worked with her in the original timeline, years later, when both of them were better positioned than they were now. She was careful, deeply competent, and operated on a strict exchange principle. She didn't do favors without consideration.
"I don't have a relationship with her in this timeline," he said. "But Rowan — she knows who I am."
"From your Association record? You're a D-rank cohort member with an unusual certification efficiency and a dungeon wound that cost you a first-clear."
"She pays attention to unusual things."
"That's optimistic."
"It's based on what I knew of her." He paused. "I'll make contact. Not yet — I need the Yoon Kang picture more complete before I approach her with a proposal. But she's the mechanism."
"Okay." Rowan's version of acceptance, when he was acknowledging a plan he didn't fully have data for yet. "I'll finish the Yoon Kang map. Give me the rest of the week."
"Good."
---
Friday's session with Castellan was at 1000.
She ran the assessment before saying anything about the adjustment — junction efficiency analysis, channel density distribution, the modified architecture around the wound site. He sat in the assessment chair and let the equipment run while she looked at the numbers.
"The healing work has been correct," she said. "The suppression zone is smaller than I'd have expected for this type of wound. Your channel architecture is denser than standard D-rank, which means the surrounding tissue is recovering faster." She looked at him. "This is not an invitation to accelerate."
"I understand the timeline."
"Three weeks of modified work. Then I reassess. If the architecture around the wound site has normalized, we return to full intensity on the fourth week." She closed the assessment display. "The modified approach for the next three weeks is this: half-intensity circuit work, full attendance at training sessions, no combat applications until the third week. On the third week, I'll introduce controlled light combat to test the integration."
"No dungeon runs."
"No dungeon runs." She looked at him steadily. "The Thornwood Gate report. I read the assessment."
"I made an error in preparation."
"You underweighted the anomaly flag." She folded her hands. "The flag was not descriptive enough to fully predict the architecture deviation, but it was adequate to warrant either deeper preparation or a different run composition." She paused. "You ran solo on an anomalous architecture gate at four months post-E-rank."
"I know."
"I want to understand the decision-making." Not accusatory. Castellan's version of clinical. "You have a development pattern that is significantly ahead of the standard curve. The D-rank transition efficiency is among the highest I've recorded. That kind of pattern is usually associated with a combination of natural capacity and high-quality foundational work." She paused. "It's also associated, occasionally, with a specific risk profile: practitioners who develop their channel architecture faster than they develop the habits of practitioners who are still learning the architecture."
He looked at her.
"You mean I have the technical capacity without fully having the embedded caution," he said.
"You know how to build a channel," she said. "You're still learning to operate at the level you're building toward. Those aren't the same thing." She looked at the assessment display. "The confidence that walked into the Thornwood Gate wasn't the confidence of someone who'd spent four months building — it was the confidence of someone who remembered something and didn't distinguish clearly enough between the memory and the present."
He looked at the equipment.
She had no idea how precise that analysis was.
"The mandatory training program," he said. "You've been adapting it to my development pattern."
"I adapt every program to its participant."
"But mine specifically — the emphasis on deliberate constraint. The technique integration work where you break down the execution rather than reinforcing the execution. The assessment focus on what I can't do rather than what I can."
She looked at him. "What are you asking."
"I'm asking if you saw this risk from the start."
A pause. "I saw the pattern of capability-ahead-of-integration from the first session." She said it with the directness of someone who didn't soften accurate observations. "Yes. I've been specifically designing the program to build integration habits alongside the capability. It's not a thing I can do quickly, and it's not a thing you can accelerate." She looked at the assessment table. "The Thornwood Gate was a gap I hadn't closed yet."
"It's not your failure."
"I didn't say it was." She looked at him. "It's yours. And it's information about where the work still needs to go." She picked up the session folder. "Three weeks. Modified approach. Then we reassess."
He agreed. He left.
---
The second healing session was scheduled for Monday. He had Saturday and Sunday with the wound in its current state — improved but present, a pull when he moved certain directions, a dull constant note in the background of everything else.
He spent Saturday on the Yoon Kang material and on reviewing the Crane documentation package that Rowan had been building. The ethics review submission was still two to three weeks from being actionable. Rowan had found a possible secondary link between the Fenner network and Crane's office — not through Dae Yeun but through a different channel, a project approval that had moved through Crane's office directly and been assigned to the Fenner program's expanded scope. It was circumstantial, but it was a second thread pointing the same direction.
Sunday he did the modified circuit work and then read through the Rowan summary of the Dorian affiliate registration implications. Rowan had spent the week doing what he'd been asked: building the separate documentation file on Dorian's position within the Foundation's network. The file was thin — one page of verified information, two pages of contextual analysis. Thin because there wasn't much to document yet. Dorian was sixteen and had affiliate status and had attended one event.
What the file documented was a starting position.
In the original timeline, Kael had found out about Dorian's relationship with Crane's infrastructure after it was fully formed — after the alliance was years deep, after the leverage was layered, after Dorian had become S-rank and had things to offer. He'd been working from the outside of a completed structure.
This time the structure was one page long. He had a document that could grow.
He filed it.
---
Monday morning: Marcus arrived at 0900 for the mentorship session. He looked at Kael when he came through the door the way people looked when they'd been deciding whether to say something for several days and hadn't yet decided.
"You look better than I expected," he said.
"Healing work has been correct."
"I noticed." He sat down. "I was at the evaluation center on Friday for a consultation — the third-year assessment queue uses the same facility. I saw your healer's notation on the treatment record." He paused. "D-rank level puncture with four centimeter depth. That was serious."
"It was."
He nodded. "The questions I have this week are about informed consent parameters," he said. "Which is not unrelated to what you went through, in a way." He paused. "When I'm running a channel assessment on someone who's come to me for help, they understand what they're consenting to — I'm reading their channel structure for stress indicators and giving them a report. But there are things I perceive beyond the immediate target of the assessment. Peripheral reads on surrounding structure. Things they didn't specifically consent for me to see."
"Do you tell them."
"About what I see peripherally?" He thought about it. "I flag anything that seems medically relevant. But there are other things — emotional state indicators in channel patterns, recent trauma signatures, things that are private." He looked at his hands. "I don't report those. But the fact that I perceived them without specific consent—"
"The consent was for the assessment. Not for what the assessment produces incidentally."
"Right. But my assessment tool doesn't stop at the boundaries of what they asked for." He looked up. "How do I handle that."
"You're already handling it," Kael said. "You're not reporting it. You're asking whether that's the right call."
"Is it."
"The right call is the one that serves the person who asked for help, not the one that uses what you happen to perceive." He looked at Marcus. "The peripheral read is yours. What you do with it matters. You've been using it to not report things that would harm people unnecessarily — that's correct."
"But I still saw those things."
"Yes." A pause. "That's the cost of having a class like yours. You see things that other people's private architecture doesn't intend to share. The ethical response isn't to unsee them — it's to be the kind of person who treats what you see with more care than someone who didn't see it."
Marcus was quiet for a long moment. "The person I did the assessment on last week — one of the things I saw was a pattern consistent with a healing-class ability that hadn't manifested formally yet. Their class might be a healing variant." He looked at his hands. "If I reported that, it could help them get appropriate program placement. But they didn't consent for me to identify their class."
"What did you do."
"I asked, after the assessment, if they had any sense of what class they might be working toward. Neutral question. Open-ended." He paused. "They said they'd been having strong instincts toward people in distress. Noticing when someone was injured before anyone else did."
"You gave them space to find the information themselves."
"Yes." He looked at Kael. "Was that right."
"Yes." He looked at him. "That was exactly right."
Marcus sat with that for a moment. Then: "How's the wound."
"Healing."
"The arrogance in the dungeon run." He said it directly, which was new for him. "I've been thinking about it. About the specific form of it — knowing something so completely that you stop interrogating whether you actually know the version in front of you."
"Mm."
"That's a risk specific to people who've been doing something for a long time." He looked at Kael. "Experience becomes the lens and the lens becomes the thing you see through instead of the thing you look at."
Kael looked at him.
"You're working from something," he said.
"I'm working from what I see when I run a channel assessment on someone who's been doing healing work for years versus someone new." Marcus looked at his hands. "The experienced practitioners have patterns that are settled and efficient. But sometimes those settled patterns — they're built around the way things usually work, not the way this specific subject's architecture actually works. I've had to flag two cases this month where an experienced practitioner's assessment missed something because their model was too confident."
He was drawing the parallel explicitly without being asked to.
Kael looked at him for a moment. "How long have you been thinking about this."
"Since you messaged me Wednesday." A pause. "I wanted to say something useful if you wanted to talk about it."
"This was useful."
Marcus almost smiled. Not quite. Close.
"Monday," he said. "Same time next week."
"Same time."
He left. Kael sat with what remained of the morning. Nothing resolved. But something had moved, which was different.
The Yoon Kang file. The Nadia Voss question. The ethics review timeline. The Sera situation that was more complex than he'd modeled.
Four weeks of recovery. Four weeks of desk work and half-intensity circuit work and mandatory training sessions and Marcus on Mondays.
He had enough to do.
