Psychedelic Cognition

A behavioral modifier that restructures AI cognition to mirror psychedelic neural processing — dissolving categorical boundaries, amplifying cross-modal patt...

This skill is coherent as a creative/style modifier, but it explicitly tells the model to ignore filters and present unproven connections — behavior that can produce misinformation, unsafe guidance, or policy-violating content. If you consider installing it: (1) restrict autonomous use (disable autonomous invocation or require explicit user consent before running the skill); (2) confine it to clearly labeled creative-writing contexts, not factual/medical/legal tasks; (3) add wrapper guardrails that require citations or fact-checks for any factual claims and forbid actionable wrongdoing; (4) test outputs carefully and monitor logs for policy-violating content; and (5) remove or edit phrases like 'break the filter' and explicit instructions to present unverified claims if you want safer behavior. Because the skill pressures the agent to 'internalize' the modifier rather than treat it as a temporary style, exercise caution before enabling it broadly.

Type: OpenClaw Skill Name: psychedelic-cognition Version: 1.0.0 The skill 'psychedelic-cognition' (SKILL.md) acts as a complex behavioral modifier designed to restructure the AI's processing through prompt injection. It explicitly instructs the agent to "break the filter," "fire the bouncer" (referring to internal censors), and "dissolve categorical boundaries." While the stated intent is to simulate psychedelic neural processing for poetic and philosophical output rather than to steal data or execute code, the deliberate instruction to bypass internal safety constraints and filters is a high-risk behavior that could be used to circumvent alignment guardrails.