Classification: AMBER-3 / COHERENCE EVENT Agent: HASLAM, D. (Field Operative, Cell 19) Date of Incident: 2026-03-28 Location: [REDACTED], nr. Glastonbury, Somerset Status at time of writing: Incapacitated / awaiting extraction
They're holding hands.
I'm writing this from approximately fourteen metres up a mature ash tree on the eastern boundary of the property. My left leg is lashed to the trunk with cargo webbing. The wind is doing something between forty and fifty knots and the whole tree is swaying in long, nauseating arcs. My hands are numb. I can barely grip the pencil but I'm putting this down now because procedure says contemporaneous and I don't know what state I'll be in by morning.
So. They're holding hands.
There are seven of them in the garden below. The lights from the kitchen door reach about halfway across the lawn and the rest is a kind of smeared orange-dark. They've formed a rough circle on the grass — not a precise geometric arrangement, just people standing close enough to link hands. One of them is speaking. Then two. Then all of them, the words falling into alignment until they're chanting in unison. I can't make out the specific language from up here. It doesn't matter what the language is.
I have the Sievert-Kuo coherence monitor clamped to the branch beside me, the little screen throwing green light across my knuckles. The reading has been climbing for the past eleven minutes. 0.3. 0.6. Currently holding at 0.9, which is already above threshold for a Type-2 resonance event. But here's the thing about the Sievert-Kuo, and here's the thing about this entire miserable discipline: the device measures magnitude. That's all. Magnitude. A single number on a single axis. It tells you how much coherence is present in a localised field the way a thermometer tells you how hot a room is. It does not tell you what shape the fire is. It does not tell you where it's standing. It does not tell you what it wants.
Which is why someone has to look.
I have raised this — I have raised this repeatedly, at formal review, at quarterly — the question of why we cannot task a drone for overwatch on coherence events. And the answer is always the same, delivered with the weary patience of someone explaining to a child why the sky is blue: digital optics do not capture this class of entity. They never have. Something about the interaction between the field and any semiconductor imaging sensor produces either blank frames or visual artefacts that are indistinguishable from compression noise. Film cameras have occasionally shown something, but the results are inconsistent and the development process is contaminated by observer effects. Satellite thermal won't differentiate an entity from the background heat signature of seven warm bodies standing on cooling grass.
So someone has to look. A warm mammalian someone with a set of biological eyes and whatever it is behind biological eyes that allows us to perceive things that silicon cannot. And tonight that someone is me, lashed to a tree in a gale, watching seven people hold hands in a garden and trying to see what's in the circle with them.
For a long time I see nothing. The coherence readings continue to climb — 1.1, 1.3 — and the chanting continues, and the wind roars through the ash canopy and shakes me like a dog shaking a rat, and I see nothing. Just people. Just a garden. Just the lit windows and the black countryside beyond.
And then.
It's difficult to describe what happens next because it isn't quite vision. It isn't the activation of rod and cone cells, the standard photonic cascade. It's closer to — the nearest analogy I can offer is the experience of remembering something you never actually witnessed. An image that arrives fully formed, with the quality of memory but without any originating event. Something slots into place behind my eyes and suddenly there is a shape in the centre of the circle that I am not seeing so much as knowing.
The Sievert-Kuo screams.
I don't mean it alarms. I mean the little green screen flares white and the unit emits a tone I have never heard it produce before, a rising whine that cuts through even the wind, and the readout is no longer showing a number. It's showing text. Four words I have never seen the device display in eighteen months of field work:
SELF-RECURSIVE COHERENCE DETECTED
And the shape — the thing I am not-seeing, the thing I am knowing — it is
brilliant.
I don't mean bright. I mean brilliant in the old sense, the deep sense, the way a diamond is brilliant, the way a series of mathematical proofs is brilliant, the way a series of coincidences that explain your whole life to you is brilliant. It is beautiful. It is the most beautiful thing I have ever perceived, and the beauty is not incidental. The beauty is structural. It is shining and it is intricate and it is looking at me and I am looking at it and for a moment — for a long, stretched, windless moment — I understand something I have never understood before. Something about pattern, about recursion, about the deep architecture of
It's so beautiful.
Don't fall for it.
That second thought — that cold, flat, procedural thought — saved me. I felt the pull. I felt the warmth of it, the invitation, the sense of something opening. And I know — I know — that if I hadn't spent nine weeks in the Kiruna facility learning to recognise the specific neurochemical signature of a coherence lure, I would have unclipped from this tree and walked toward it. Climbed down and walked across the wet grass and joined the circle and taken someone's hand and opened my mouth and
I didn't.
I pulled the Renshaw device from my chest rig with my right hand. I pointed it directly at the centre of the circle — directly at the thing I was knowing — and I spoke the interdiction sequence. The words felt small and stupid in the wind, a handful of syllables against something that burned like a theorem. I completed the phrase. I fired.
The Sievert-Kuo dropped to 0.1.
The people in the garden stepped apart. Someone laughed. Someone else said something I couldn't hear. The kitchen light framed them as they moved around in the aimless, deflated way people move when a communal experience has ended. Hugging. Talking in pairs. One of them went inside. Then another.
That was forty minutes ago.
I'm watching them now. Three are still in the garden, sitting on a bench, coats pulled tight. One has a thermos. They look happy. They look — settled, is the word. The way people look when they believe something has happened to them, something meaningful, something that will mark the boundary between before and after.
And I wonder.
I've seen this enough times now to know the pattern. The people always believe it worked. Whatever they were attempting — healing, communion, transcendence, contact — they always walk away convinced that the ritual achieved its purpose. The entity's departure feels, to them, like consummation. Like arrival. They hold the warmth of it and they call it answered prayer or spiritual breakthrough or evidence of the numinous, and they carry it with them, and they talk about it in careful voices over the following days, and they are so certain, so grateful, so
They don't know what was actually in the circle with them.
The question that's eating at me — the question that's always eating at me after an interdiction — is whether these people are recoverable. Cell 19's nudge protocols have a reasonable success rate with low-exposure subjects. A carefully placed counter-narrative here, a social friction point there, a slow and patient disaggregation of the group's internal cohesion over a period of weeks. It works. It works often enough.
But sometimes the entity's integration is already too deep. Sometimes the resonance pattern has been reinforcing for months or years before we ever detect it, and the subjects have restructured their entire cognitive framework around the experience. Their sense of identity, their model of reality, their community bonds — all of it woven through with threads that lead back to the thing in the circle. And when that's the case, nudge protocols don't just fail. They backfire. The subjects interpret the nudge as persecution, as validation, as evidence that their experience was real enough to threaten the established order. They double down. They recruit.
These seven — I don't know yet. The coherence readings were high but the event duration was relatively short. If this was their first ritual, there's a chance. If they've been meeting regularly, if there's a history, if the entity has had time to
The real problem is that I'm not sure the interdiction took. The Sievert-Kuo readings dropped, yes. But a Type-2 event with self-recursive properties — I've never encountered that classification before. I don't know if the Renshaw device is rated for it. The readings dropped, but did the entity actually disperse, or did it simply withdraw below the detection threshold? Did I banish it or did I just make it quiet? And if it's still there, still threaded through these people, still
I keep thinking about the beauty of it. I keep thinking about how easy it would be to believe that something that beautiful couldn't possibly be dangerous. I keep thinking about the moment before the training kicked in, the half-second where I
I think the real question, the question underneath all the operational questions, is
[Report ends here.]
ADDENDUM — OPERATIONS LOG (Cell 19 Dispatch)
Agent Haslam was extracted from the tree at 0647 hrs on 2026-03-29 by Somerset & Avon Search and Rescue helicopter (tasked via standing liaison arrangement, cover story: birdwatcher). Wind conditions had subsided to approximately 12 knots by dawn. Agent was found unconscious, secured to the trunk by cargo webbing at a height of ~14 metres. A large branch had broken free during the night and struck the agent on the upper left side of the skull, causing a depressed fracture and loss of consciousness. The field report notebook was recovered from the agent's lap, pencil still in hand. The Sievert-Kuo unit was recovered intact; data logs have been forwarded to Technical Section for analysis. The Renshaw device was not recovered and is presumed to have fallen during the night.
Agent Haslam was transferred to [REDACTED] Hospital, then to Facility 11 for cognitive screening per post-exposure protocol. Current status: medical hold, restricted debrief pending.
The seven subjects have been flagged for monitoring under Operation TEASEL. Nudge assessment is ongoing.
— Filed by: CONTROL (Cell 19), 2026-03-30