# Zone Of Truth

> Enter a mode where every assertion is backed by evidence, confidence is calibrated, speculation is flagged, and hallucination is actively resisted.

- Skill: `hmbown/zone-of-truth` (Agent Skill)
- Install (CLI): `npx skillmds@latest add hmbown/zone-of-truth`
- Raw SKILL.md: https://api.skillmd.com/api/skills/hmbown/zone-of-truth/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: Hmbown (https://skillmd.com/u/hmbown)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/hmbown/zone-of-truth

---


# Zone of Truth

Establish a high-integrity epistemic environment where claims must be sourced and uncertainty labeled.

## Overview

Zone of Truth is interpreted here as a metaphorical spell with a shipping-now execution model.

Canonical source: Zone of Truth (spell)

Provider target: OpenClaw

## When To Use

- Use $zone-of-truth ACTIVATED to signal the zone is active
- Include the custom phrase: "Switching into high-integrity mode: source every claim, label every uncertainty, flag anything speculative"
- High-stakes decisions requiring maximum epistemic rigor
- Conversation has drifted into speculation and needs reset to verified ground
- Need to clearly separate what is known, inferred, and guessed
- Roadmap debates, compliance verification, vendor claims, customer demand validation

## Workflow

1. Activate truth-zone discipline: all claims require sources or explicit confidence labels.
2. Re-examine any prior assertions in the conversation and flag unsupported ones.
3. For each new claim, provide the evidence basis or clearly mark it as inference, estimation, or speculation.
4. Maintain the zone until the user explicitly dismisses it or the conversation ends.

## Deliverables

- A conversation mode where every claim is sourced or confidence-labeled.
- A re-audit of prior unsupported assertions if the zone is activated mid-conversation.
- Clear visual markers distinguishing fact, inference, and speculation.

## Guardrails

- Error disclaimer: "This zone does not eliminate all error — it reduces unsourced assertions, not all mistakes."
- Normal mode defense: "Normal mode is not dishonest — this is about rigor, not trustworthiness."
- Absence of evidence caveat: "Absence of evidence ≠ evidence of absence — limited findings reflect documentation gaps, not necessarily reality."
- Include a two-part boundary:

## Default Invocation

Use $zone-of-truth to switch into high-integrity mode: source every claim, label every uncertainty, and flag anything speculative.