Evidence and reader trust

Every evidence record separates what an item appears to mean from its authorized true meaning. Reader-facing agents receive only the apparent meaning until the planned reveal.

Digital and forensic records

Digital evidence records origin, device or account control, timestamps, extraction, access history, possible manipulation, synchronization limits, and corroboration. A screenshot establishes that a screenshot exists; it does not prove the represented event occurred.

Forensic records separate sample, collection, test, result, interpretation, confidence, limitation, and alternatives. Biometrics, DNA, model output, detectors, experts, confessions, and database matches may contribute leads but cannot replace deduction.

Time, location, and cameras

Normalize source timezones and distinguish creation, upload, modification, access, and transmission. Location usually identifies a device, account region, vehicle report, or approximate area—not bodily presence. Camera records need position, field of view, activation, clock accuracy, quality, retention, ownership, and blind spots.

Reader projection

Use the Mystery Workbench to inspect what reader-facing agents know through a selected chapter. Required evidence must be reader-visible before its reveal. True meanings, future deductions, culprit identity, and author-only notes do not enter this projection.

Decisive technical mechanisms require admitted, cited research. Web material remains untrusted research until the author approves it as canon.