How to make theft detection powder in 5 minutes, and the one defect every recipe shares.

The cornstarch and oil paste. Mix two parts cornstarch with one part light mineral oil (or food-grade petroleum jelly) and stir until uniform. Optional: a pinch of talcum powder to lift the bulk and a pinch of fluorescent pigment if you want it visible under UV. Spread a thin film on the inner shelf of a mailroom drawer, the underside of a copper spool flange, the top edge of a tool crib lid, or the back of a cash drawer divider. The mixture transfers easily to dry skin, lasts a few days through casual hand washing, and costs under five dollars per ounce. The recipe is fine. What it lacks is any way to defeat a defense lawyer who asks when the suspect's hand contacted the powder.

Yes, three ways. First, fluorescent pigment (sold by the gram on Amazon as 'invisible UV powder' or 'glow-in-the-dark pigment 365 nm') mixed at roughly 1:10 with cornstarch or talc, then suspended in petroleum jelly if you need it to stick to vertical or outdoor surfaces. Second, a quinine slurry from tonic water: reduce 250 ml of brand-name tonic water (Schweppes, Canada Dry, Fever-Tree all contain quinine) by simmering to one-tenth volume, then mix the residue with cornstarch and let it dry; the dried powder fluoresces blue under a 365 nm UV flashlight. Third, riboflavin (vitamin B2) crushed from a tablet and dusted into talcum powder fluoresces yellow-green under 395 nm. All three transfer to skin and survive at least one hand wash. None of them carry a verifiable batch number, which matters in court more than the formula does.

Commercial visible-stain powders (banker dye packs, SmartWater visible stains, the purple amino-acid-reactive powders sold by ASR Federal and J&L Self Defense) are formulated to bind covalently to skin proteins and stay visible for days through repeated washing. A homemade fluorescent pigment in petroleum jelly is a mechanical transfer only: it sits on the skin, it washes off in one or two hand-washings with soap, and it does not penetrate clothing fibres in the same way. For deterrence and recurring-thief identification within the same shift, the homemade version is acceptable. For 'still visible at the police station three hours later,' the commercial protein-binding powder is materially better. Either way, the camera-side record that timestamps the contact is the same, and the same record is what makes either one usable as evidence.

Three defenses, all routinely successful. Secondary transfer: the suspect's hand picked up the powder from a door handle a legitimate user touched first, so glow-on-hand proves contact with powder, not contact with the stolen item. Similar fluorescence: many household substances glow under UV (laundry detergent residue, toothpaste, tonic water itself, petroleum jelly, olive oil, some bodily fluids), so the prosecutor needs to prove the glow on the hand is the same powder the building applied. With a commercial powder, a lab can match the batch number against the manufacturer's spec. With a homemade powder, there is no batch number; the prosecutor has to rebuild the recipe in a forensic lab and demonstrate spectral identity. Chain of custody: when was the powder applied, who applied it, when was the hand inspected, by whom, with what UV wavelength. All three defenses collapse the moment the prosecutor can pair the powder transfer with an independent timestamped record of the suspect physically standing over the treated surface at a specific second. Without that record, every homemade powder case is a coin flip.

A pre_action_zone_entry event from a Cyrano edge unit. The event includes a tile thumbnail (480x270 crop centered on the person), the zone name (a polygon you draw directly over the treated surface), dwell_seconds (how long the centroid stayed inside the polygon), an ISO 8601 timestamp with timezone offset to the second, an overlay_mask field that records which DVR-baked overlays were masked before inference (the clock, the camera-name strip, the channel bug), and a delivery channel (a property's WhatsApp ops thread). Capture-to-WhatsApp latency is under 60 seconds. The event is the independent corroborating record that puts the suspect at the treated surface at a specific moment, which is the one thing a homemade powder transfer cannot do on its own.

No. The capture point is the DVR or NVR's HDMI multiview output, the same signal that drives the existing guard monitor. The composite already has every camera mosaiced into tiles, so a single HDMI tap gives inference access to up to 25 feeds at once. Works on analog cameras, IP cameras, and any DVR brand that has HDMI out: Hikvision, Dahua, Lorex, Amcrest, Reolink, Uniview, Swann, Night Owl, Q-See, ANNKE, and their rebrands. No camera firmware involvement, no ONVIF negotiation, no per-camera credentials. Physical install on a running DVR is under 2 minutes: HDMI in, HDMI out to the monitor, network, power. Inference happens on the edge unit, on the multiview composite. The DVR you own does not move.

Three settings. Mailrooms and parcel lockers where the recurring loss is low-dollar but constant; a homemade fluorescent paste on the inner shelf, paired with a 19:00 to 06:00 zone window, identifies the recurring thief inside a week. Tool cribs and gang boxes on construction sites where the powder is consumable and gets reapplied weekly anyway; cornstarch + UV pigment is a fraction of the cost. Cash drawers and till trays during shift change; a thin smear of homemade visible-stain powder on the underside of the cash tray, paired with a tight zone armed only at shift change, narrows substitution and skim cases to a short list of suspects. Where commercial wins: anything you might actually litigate, where a verifiable batch number and protein-binding chemistry pays back the cost.

Date, time, applicator initials, exact surfaces treated, recipe (with measured ratios), the source of any pigment or quinine reduction, the wavelength of the UV flashlight you intend to use for inspection, and the re-application cadence. For a homemade recipe, the log is doing double duty: it is both the chain-of-custody record and the substitute-for-batch-number that lets a forensic lab verify the powder later. Pair it with the zone configuration record (which polygon, which camera tile, what dwell threshold, what armed window) and the two together form the front half of the evidence chain. The zone entry event and the UV inspection record form the back half.

Yes, and it is the entire point of pairing the two. The Cyrano alert lands on WhatsApp under 60 seconds after the zone fires; if the property has a front-desk attendant or a contracted patrol, the suspect is often still on site or in the immediate area. A 365 to 395 nm UV flashlight inspection in the next several minutes catches the glow at peak intensity. Photograph hands, sleeves, and any held items. Log the wavelength of the UV flashlight used, the time of inspection, and the inspecting officer. Save the photographs alongside the zone event payload. That single bundle is the correlation exhibit that a prosecutor needs to defeat secondary transfer and similar-fluorescence challenges in one motion.