Baseline- Part Six

Part Six — Records

Records Support is quieter than intake.

No residents. No voices. Just terminals arranged in neat rows, each assigned a narrow scope of access. The air feels different here—cooler, steadier. Like a place designed to slow the heart.

I like it at first.

My task is verification. Cross-checking archived data against current models. Flagging inconsistencies. Correcting clerical drift. It’s careful work. Necessary work. The kind that keeps systems honest.

I pull up a completed subject file at random and overlay it with the current baseline template. The curves don’t quite match. Minor variance. Expected. Bodies differ.

The next file is closer.

The third makes me pause.

The baseline curve—heart rate variability, stress response, recovery slope—is identical to mine.

Not similar. Identical.

I check the subject designation. Different. I check the date stamp. Years old. I compare it again to my staff health profile, pulled automatically from occupational monitoring.

The lines sit perfectly on top of each other.

That shouldn’t happen.

I flag it as a duplication error and open a correction request. The system processes for a moment longer than usual before returning a response.

Access denied: insufficient clearance.

I try again, assuming I’ve mistyped the request code. Same result.

That’s unusual. Clerical errors fall well within my scope. I scroll back through the file, looking for the lock trigger. That’s when I notice the category header.

Subject Status: Conversion Outcome.

I’ve never worked with those files before.

I attempt to open the metadata.

Access denied.

There’s a faint irritation now, like grit under the skin. I don’t recognise it immediately. Records aren’t meant to provoke feeling. They’re meant to resolve it.

I run a wider search, filtering for baseline curves within a narrow tolerance band. The system returns multiple results. Too many.

I select one at random.

Identical.

Another.

Identical.

Each file marked completed. Each archived. Each flagged as successful.

I shouldn’t be able to see this much. The fact that I can suggests the system expects me to.

Still, when I try to open the earliest file, the screen freezes briefly before the denial appears again.

Insufficient clearance.

I lean back in my chair and wait for the irritation to pass. It doesn’t.

At the end of the shift, I collect the printed reports from the output tray. Most are routine summaries. One page sits apart from the rest, slightly misaligned, as if it was never meant to be there.

I almost put it back without looking.

The header catches my eye.

Baseline Comparison — Archived Dataset.

My staff ID is printed at the top, listed under “Reference Profile.”

I read it twice, then a third time.

Below it is a subject designation. Redacted. Date stamped far earlier than my employment record begins. The baseline curve printed beneath is the same one I’ve seen all afternoon.

Mine.

I look around the room. No one is watching. The other terminals hum softly, indifferent.

I fold the page once and slide it into my pocket.

Clerical errors happen. Misprints happen. Systems overlap.

There will be an explanation.

There has to be.

I log the anomaly as unresolved and sign out for the day.

As I leave Records Support, the system updates my task queue for tomorrow.

Further archival review assigned.

Priority: high.

Baseline integrity must be maintained.