field guides to manipulation patterns
Clear anatomy of social engineering, manufactured urgency, framing traps, and synthetic trust cues—written plainly for both human readers and agent builders.
no hacker-movie glitch theater; calm eyes beat panic.
cognitive security · person and bot
Influence operations, social engineering, persuasion tricks, and prompt-level attacks on AI agents all share one target: how decisions get steered without you noticing.
cogsec.guru names the trick plainly, maps how the lure tries to slip past judgment, and keeps signal teal reserved strictly for clarity—never letting bait red touch the eye.
what cogsec.guru examines
Clear anatomy of social engineering, manufactured urgency, framing traps, and synthetic trust cues—written plainly for both human readers and agent builders.
no hacker-movie glitch theater; calm eyes beat panic.
Inspecting how indirect injections, hidden instructions in retrieved pages, and subtle context nudges attempt to hijack an agent’s decision loop.
three hooks come in; none reach the keyhole iris.
Every motion snaps on a deliberate beat—one 30-degree gear tooth turn, a signal teal flash, and a steady open eye before settling cleanly back to rest.
red always loses at the perimeter.