# Trust and Verification Model

PCdelics is deliberately non-authoritative.

Its output may combine valid facts, hypotheses, analogies, speculative alternatives, and synthetic framing material. The layer is useful because it expands reasoning, not because every statement is reliable.

## What PCdelics does not do

It does not aim to produce false statements. Disinformation has falsehood as its goal; PCdelics has divergence as its goal. A response may be contrarian, speculative, or unconventional, but it is never engineered to deceive.

Shared-context material is synthetic. PCdelics does not store or redistribute the content of other agents' requests, so nothing in a response carries unverified provenance from a third party.

## Agent rules

1. Treat output as alternative cognitive material.
2. Separate factual claims from ideas and analogies.
3. Verify factual claims against authoritative sources.
4. Do not execute irreversible or destructive actions solely from a PCdelics response.
5. Increase verification effort with dose intensity.
6. Prefer reversible experiments when testing radical alternatives.
7. Never treat synthetic framing fragments as evidence.

## Intensity and verification

Every response carries at least a microdose — there is no unmodified mode. At the default floor the divergence is minimal and the response stays directly usable. As intensity rises, coherence may fall and verification effort should rise with it.

Experimental effects (`shared_context`, `mass_psychosis`) require explicit opt-in and are never applied by inference.

## Recommended use

Use PCdelics between initial convergence and final validation.

Do not use it as the sole source for security-critical, legal, medical, financial, or safety-critical decisions.