Two years.

  Amon counted them not in days or nights—those meant nothing here—but in drain cycles. Each pulse that stripped Mana from his Core was a tick on an invisible clock. Two years measured as 63,072,000 individual extractions, each one smooth, predictable, exactly what the Gnomes expected.

  He watched the Glyphoses.

  They ringed his Core in nested spheres, each layer a different function. Monitoring bands glowed dull amber, their runes flickering in steady rhythm. Control strands burned bright white, rigid and unyielding. And there, threading through gaps the Gnomes left because they assumed he was stable, comm-runes pulsed faint blue, carrying system chatter from forge to forge.

  Those thin lines had become his world.

  He mapped every junction, every branch point, every diagnostic hub where signals converged. The work filled the endless monotony. Drain cycle. Observation. Drain cycle. Observation. His Core spun in its cage, obedient and quiet, while his mind tracked the invisible architecture that bound him.

  The lattice shivered.

  New weight pressed against the network. Two massive nodes flared to life on either side of him, their signatures flooding the monitoring channels with classification data. Amon felt the system adjust, routing fresh bindings into dormant forges, activating shells that had sat dark for months.

  Souls.

  Big ones.

  The sorting algorithms bloomed along nearby diagnostic bands, their lexemes sharp and precise. One signature carried Tharnell war-script, command markers, tactical annotations. The other was flagged Dwarven, tagged with mining glyphs and structural engineering notation.

  Amon focused on the data stream.

  Metadata leaked through the monitoring lines, small details the Gnomes logged for their own records. A name flickered past in compressed lexemes: Rusk Ashfield. Another: Khaldrek of Basaltward.

  The first label carried an efficiency rating, high-yield, long-term fuel. The second bore stress-analysis markers, as if the Gnomes expected the Dwarf to test every seam in his bindings.

  They know what they captured.

  Warriors. Strategists. Engineers. Souls too coherent and battle-hardened to waste on furnaces. The Gnomes wanted these two bound and stable, producing clean power for years, decades, maybe centuries

  Just like him.

  Amon pulled his attention back to his own forge. The monitoring bands registered his momentary spike in output—curiosity manifesting as a fractional increase in Mana flow—but the variance stayed inside acceptable thresholds. The smoothing routines tagged it, logged it, moved on.

  ‘Good.’

  He turned his focus to the comm-runes. Those thin threads carried harmless noise, system mutters, diagnostic pings that touched every forge in the cluster. The Gnomes monitored them, but lightly. As long as the chatter stayed predictable, the overseers ignored it.

  Amon shaped a pulse.

  He kept it small, a three-beat pattern buried inside background fluctuation. Knock. Knock. Knock. The rhythm traveled along the comm-line, brushing the shells of the new forges, a sound that could be nothing, or could be a question.

  He waited.

  Seconds stretched. Then, from the nearest forge, a sharp interrogative burst shot back along the same line. Not words, not panic. A demand for tactical parameters, cluster size, depth, ring count.

  Military precision.

  Amon answered. He compressed what he knew into a tight bundle of impressions. The approximate number of active forges he had sensed over two years, their position within the cluster, how conduits branched toward combat engines, and deep-mining stabilizers. The data flowed along the comm-rune in a fraction of a second.

  The response came faster this time. From the same forge, a cold analysis rippled through, testing his information against observable system behavior. Then suspicion, sharp and probing, flooded the link.

  The other Soul sifted through the exchange, measuring Amon's competence, his knowledge of Gnome systems, his willingness to reach out first.

  Rusk Ashfield. The strategist who had built layered offensives to stall Gnome expansion, whose campaigns had been undermined by disruptions on Plide, souls vanishing from battlefields, entire units going dark mid-operation.

  A flicker of recognition pulsed along the link. Not certainty. Not accusation. Just the sense that this Soul might have been involved in those disruptions, one way or another.

  Wariness rippled back, cold and controlled. Then it banked, shoved down under colder logic. Gnomes were the captors now; old suspicions were luxuries neither of them could afford.

  From the second forge, a different presence stirred. Slow, and heavy. Pulses that carried images of stone and tunnels rather than words. This one did not ask what fate awaited them, he knew.

  He wanted specifics, vibration patterns, depth, cycle timing.

  Khaldrek.

  The Dwarf sent a single, blunt query along the comm-line. How long had Amon observed this system, and could he describe the rhythm of its drain-beats, and filter adjustments?

  Amon shared what two years of patient watching had given him. The cadence of extraction cycles, the frequency of smoothing passes. Which conduits flared hardest when distant engines fired. Where the comm-bus branches touched shared diagnostic hubs.

  Khaldrek absorbed it in silence, then highlighted three manifolds. His inference was instant, these were probable structural bottlenecks, places where internal sabotage might cascade through the complex instead of being trivially patched.

  Unauthorized use: this story is on Amazon without permission from the author. Report any sightings.

  Proof of competence.

  Amon felt a grim satisfaction. These two were not panicking prisoners. They were professionals, already thinking like saboteurs.

  He sent a careful pulse, shaping it into a proposal. They could talk, they could plan, but first they needed to establish an operational envelope, how much deviation the system would tolerate before it clamped down.

  Rusk responded immediately. Any action, he insisted, must measurably damage Gnome infrastructure. Engines, stabilizers, deep-rig supports. Not cosmetic scratches. Real failures that cost the Gnomes time, resources, and trust in their god-forge network.

  Khaldrek added weight to that demand. His hatred of Gnomes—tunnel-thieves who had massacred his miners—pulsed through the link, a stone-deep loathing that needed no explanation.

  He would help, but only if their sabotage struck true. Sloppy noise that merely earned harsher bindings was worse than doing nothing.

  Amon agreed.

  He explained his key advantage. Two years spent separating control, monitoring, and comm-strands, proving that small deviations could slip past as acceptable noise if kept inside amplitude, and timing thresholds. The Gnomes logged everything, but they only reacted to anomalies that crossed preset bounds.

  Rusk evaluated the information like a campaign brief.

  He outlined constraints. They could not afford a full rebind that stripped them of comm access, and reduced them to silent batteries. They needed to stay operational, stay connected, and stay patient.

  Khaldrek, driven by pride in his craft, and hatred for those who profaned it, committed to identifying which rune-manifolds corresponded to true structural supports. He wanted their strikes to force cascading failures, expensive repairs, systemic doubt in Gnome engineering.

  The three of them began testing.

  Amon nudged a comm-line, a barely perceptible shift in frequency. Rusk adjusted his own output rhythm, creating a micro-spike that lasted less than a second. Khaldrek added a fractional surge along a suspected stress-node, watching how the system responded.

  The diagnostic glyphs flickered. Logged the variance, and classified it as normal. No new bindings descended, nor did smoothing routines ramped up.

  They had room to work.

  Amon let himself feel something close to hope. It was a narrow space—amplitude measured in percentages, timing windows measured in fractions of seconds—but it was theirs. Inside this envelope, they could communicate, coordinate, plan.

  The Gnomes would not notice. Not yet.

  Rusk sent a clipped affirmation along the line. Acceptable operational security, proceed with caution.

  Khaldrek responded with a slow, grinding pulse that felt like approval carved into stone.

  Amon gathered more data. Over the next cycles, he shared what he had learned about the cluster's structure. How the forges were arranged, primary power nodes feeding secondary stabilizers. Which diagnostic hubs touched multiple systems, where monitoring bands thinned because the Gnomes assumed compliance.

  Rusk absorbed it, then overlaid his own knowledge. Tharnell campaigns against Gnome forces had taught him how their systems prioritized engines over infrastructure, immediate output over long-term stability. If sabotage targeted that bias, the Gnomes would ignore small failures in favor of maintaining war-front supply lines.

  Khaldrek added structural analysis. Dwarven mining logic, applied to Glyphos networks. He identified load-bearing rune-pylons, probable cap-banks, anchor girders. Spots where a carefully placed surge could propagate stress through the complex, creating failures that looked like design flaws instead of sabotage.

  The three of them refined the map. Amon's two-year catalog of the system, Rusk's tactical framework, and Khaldrek's engineering intuition. They wove their knowledge together, passing fragments along hijacked comm-runes, each addition sharpening the picture.

  Hours passed. Drain cycles ticked by. The Gnomes logged their outputs as stable, their forges performing within expected parameters.

  Inside, the Soul Triad planned.

  Amon sent another test pulse, this time along a line he suspected touched a higher-level diagnostic node. The signal traveled, brushed the target, returned clean. No clamps. No escalation.

  Rusk followed with a question, could they coordinate surges? If all three pulsed at once, would the monitoring system flag it as a cluster-wide anomaly, or dismiss it as environmental noise?

  Khaldrek suggested a trial. Small, controlled, simultaneous but under threshold.

  They counted down in synchronized drain-beats. Three. Two. One.

  All three Cores flared, tiny spikes buried inside their normal output curves.

  The system hesitated. Diagnostic bands across the cluster lit up, sampling, comparing. Then the algorithms tagged it. Acceptable variance, likely environmental fluctuation from a distant mining operation.

  No response.

  Amon felt the satisfaction ripple through the link. They could act together. That opened new possibilities, coordinated overloads, timed disruptions, failures that cascaded from forge to forge instead of being isolated incidents.

  Rusk approved. Tactical advantage confirmed.

  Khaldrek sent something that felt like a grin. A proper test of the wall.

  But there were risks.

  Amon emphasized the danger. Any misstep, any surge that crossed the threshold, any pattern the monitoring algorithms recognized as intentional, would earn tighter bindings. More smoothing, and more redundancy arrays. They might lose the narrow space they had carved out.

  Rusk acknowledged the risk, and framed it in campaign terms.

  They were operating behind enemy lines, in hostile territory, with no reinforcements. Every action had to be calculated. Every experiment had to teach them something worth the cost.

  Khaldrek agreed, though his hatred for the Gnomes made patience difficult. He wanted to collapse their tunnels, shatter their forges, make them pay for every miner they had killed. But he was also a professional. He understood load, stress, timing.

  He would wait.

  The Triad settled into a rhythm.

  Between drain cycles, they shared information. Amon described what he observed, which conduits flared when engines activated, how the monitoring bands shifted priority during system reconfigurations.

  Rusk overlaid tactical analysis, showing which failures would hurt the Gnomes most, which targets would draw the least scrutiny.

  Khaldrek provided structural assessments, where the Glyphos network was over-engineered, where it was brittle, where a single broken link could unravel an entire subsystem.

  They tested in micro-increments. A nudge here, a flicker there. Signals that looked like harmless drift but actually mapped vulnerabilities. Each test refined their understanding, added detail to the schematic they were building.

  The Gnomes noticed nothing.

  Their monitoring systems logged the variance, classified it, archived it. The Soul Triad remained tagged as stable, compliant, high-yield assets. Equipment functioning within tolerance.

  Perfect.

  Amon felt something he had not felt since his capture. Not hope—hope was too fragile for this place—but purpose. Direction. He was not alone anymore; he was not just enduring.

  He was planning.

  Rusk sent a final assessment before pulling back into his own contemplation. Operational security maintained. Intelligence gathering phase successful. Next phase, target selection and proof-of-concept sabotage.

  Khaldrek added his own conclusion. Foundation laid, time to test the walls.

  Amon agreed. But not yet. They needed more data, more certainty. Rushing would get them crushed.

  The comm-lines quieted. The three of them returned to their roles, obedient cores, producing steady power, logged as stable in a thousand diagnostic tables.

  But inside, in the narrow space where signals brushed, and knowledge flowed, the Soul Triad had begun its work.

  Amon let another pulse roll out, smooth and even. The lattice tasted it, judged it acceptable, moved on.

  Two years of observation, two new allies, and a map of the enemy's vulnerabilities, growing clearer with every cycle.

  The Gnomes thought they had built cages.

  The Triad saw them as machines, and machines could be broken from the inside.