All Anomaly Types in Night At The Infirmary

Night At The Infirmary asks you to inspect human patients through three verification layers — the reception window, a printed photo, and CCTV security feeds — before admitting anyone into the infirmary. Anomalies are disguised entities that fail one or more of those layers. This page is the master index: every documented tell category from beta community reports under UPD 1, grouped by where you spot it and what action to take.

Anomaly types are not always separate monsters with fixed names. Many are tell families — patterns like hollow eyes, photo mismatches, or void-black CCTV silhouettes — that can appear on different patient models across nights. What matters is the layer they appear on and whether the mismatch is clear enough to justify hitting the shutter reject button. When in doubt, cross-reference the dedicated pages for visual tells, photo tells, and camera tells.

Use this index alongside the interactive checklist during play. The checklist enforces layer order; this page explains what each category means when you spot something unusual.

Visual anomalies expose themselves when you inspect the live patient through the reception window. These are often the fastest tells to spot because they require no equipment — only careful observation before the patient advances to the photo booth.

• Hollow eyes — Empty or sunken eye sockets that do not reflect light normally. Often paired with unnatural stillness.
• Wrong eye count or spacing — Three eyes, misaligned pupils, or eyes placed too wide or narrow for a human face.
• Sharp or excessive teeth — Grins that show teeth too long, too many, or shaped unlike human dentition.
• Proportion errors — Head too large or small for the body, arms that hang at wrong angles, or height that does not match voice cues.
• Unnatural movement — Twitching limbs, jerky head turns, or complete frozen stillness while the patient claims to be nervous.
• Head tracking — The patient's head follows your movement at the window even when their body stays fixed, suggesting wrong awareness.

Full detail and scanning order live on the visual tells page. If any visual tell is confirmed, you can reject immediately — but beta veterans still run photo and CCTV when time allows, because some window-clean patients fail later layers.

Photo anomalies appear only when you compare the printed desk snapshot against the live patient. The photo booth captures an identification still that should match the person in front of you one-to-one. When it does not, you have a photo-layer anomaly regardless of how normal the window check felt.

• Feature mismatch — Eyes, ears, nose, mouth, or hair in the print differ from the live patient. Classic one-to-one comparison failure.
• Added or removed traits — Horns, extra eyes, or scars visible on the print but absent at the window, or the reverse.
• Static photo — Heavy grain, TV snow texture, or corrupted image data on an otherwise normal-looking print.
• Cursed photo — Bloodshot eyes and an unnatural grin on the print. Viewing can drain sanity; reject without staring once confirmed.
• Hyper-realistic face — The print shows skin detail or expression more lifelike or wrong compared to the patient standing at the desk.

See the complete comparison workflow on photo tells. Never skip this layer because the window looked clean — photo-only skinwalkers are among the most common early-shift deaths in beta reports.

Camera anomalies exist only on the security monitor feeds. A patient can pass the window and photo checks yet reveal a completely different form on CCTV — stretched limbs, void-black bodies, or a skinwalker shape invisible in person. Always cycle through entrance, hallway, and waiting-area cameras before admitting.

• Black box eyes — A rectangular void covering the eyes on feed while the face looked normal at the window.
• Void / pitch-dark body — Silhouette entirely black on camera with no visible clothing detail.
• Stretched or distorted limbs — Arms, legs, or neck elongated beyond human range on the feed.
• Camera stare — Patient looks directly into the CCTV lens while walking, not matching natural eye behavior.
• Feature mismatch vs window — Different ears, nose, or jaw on feed compared to what you saw in person.
• Skinwalker form — Wrong shape visible only on camera; may include non-human posture or extra appendages.
• Sanity events — Feed zooms toward a dark figure with red eyes. Look away and switch cameras; prolonged viewing drains sanity.

Feed-specific guidance is on camera tells. Re-open feeds for a second look when unsure — some anomalies re-roll camera presentation between switches.

Some patients carry tells on more than one layer. A visitor might show subtle wrong teeth at the window, mismatched ears on the photo, and a void body on CCTV — each alone would justify rejection, and together they remove any doubt. Other entities deliberately hide on one layer only: normal at the window, cursed print, clean CCTV, or any other combination that punishes skipped steps.

Night difficulty in UPD 1 beta scales tell frequency and subtlety. Early nights favor obvious visual errors; later shifts introduce camera-only skinwalkers and static photos on patients who speak and stand like humans. The all nights walkthrough maps when new families tend to appear so you can pre-read the right subpage.

Decision rule for every type on this list: reject on the first confirmed tell on any layer. Admission requires agreement across window, photo, and CCTV. For procedural help, read how to detect anomalies and how to reject patients. Return to the anomalies hub for links to every subguide.