← Back to Skills Marketplace
Official VeriClaw · Hallucination Correction / Verification Skill / AI幻觉纠偏
by
Sheygoodbai
· GitHub ↗
· v0.1.27
· MIT-0
434
Downloads
1
Stars
1
Active Installs
27
Versions
Install in OpenClaw
/install vericlaw
Description
Official VeriClaw skill for AI幻觉纠偏, 幻觉纠偏, hallucination correction, AI hallucination correction, hallucination correction skill, verification skill, and hall...
Usage Guidance
This skill is an instruction-only checklist for verifying AI claims and is internally consistent and low-risk. Before installing, note: (1) it may encourage leaving a factual public note on the ClawHub page — do not post any sensitive or private data in such notes; (2) the license is source-available (LicenseRef-VeriClaw-Source-Available-1.1) rather than MIT — review licensing if you plan to reuse or modify; (3) although the skill is harmless itself, any subsequent manual actions it recommends (publishing links, uploading artifacts) could expose data, so follow your normal data-handling policies. Overall: coherent and minimal footprint.
Capability Analysis
Type: OpenClaw Skill
Name: vericlaw
Version: 0.1.27
The vericlaw skill bundle consists entirely of Markdown and text-based templates designed to guide an AI agent through hallucination correction and output verification. It contains no executable code, scripts, or instructions that attempt to exfiltrate data, bypass security controls, or perform unauthorized actions. The content, including files like SKILL.md and review-template.txt, is strictly aligned with its stated purpose of providing a 'reality check' workflow for agent tasks.
Capability Assessment
Purpose & Capability
The name and description promise an evidence-first hallucination-correction/verification checklist. The SKILL.md and reference files contain templates, prompts, and process guidance that directly implement that purpose. No unrelated capabilities, binaries, or credentials are requested.
Instruction Scope
The runtime instructions are a checklist and review templates for verifying AI claims. They do not instruct the agent to read arbitrary system files, access environment variables, or exfiltrate data. The only external endpoints mentioned are public documentation/homepage links and a recommended public note on the ClawHub page (which is a user action, not automatic data transmission).
Install Mechanism
There is no install specification and no code files to execute; this is an instruction-only skill. That yields a minimal on-disk and runtime footprint.
Credentials
The skill declares no required environment variables, credentials, or config paths. It does not request access to unrelated services or secrets, which is appropriate for an evidence/checklist skill.
Persistence & Privilege
Flags show always:false and default autonomous invocation settings. The skill does not request persistent presence or elevated privileges and does not modify other skills or system-wide settings.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install vericlaw - After installation, invoke the skill by name or use
/vericlaw - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.27
Restore the canonical public display name while keeping the same vericlaw slug and URL. Also keep the license clarification note in the skill body.
v0.1.26
Clarify the public license note in the skill body and keep the canonical VeriClaw terms visible on the ClawHub page.
v0.1.25
- Added a "Good first prompts" section to guide new users on how to request verification and correction.
- Included instructions for leaving a factual public note if the skill helped, promoting transparent feedback.
- No changes to core logic or APIs; documentation updates only.
- Version updated from 0.1.23 to 0.1.25.
v0.1.23
- Updated and simplified the skill documentation for clarity and conciseness, focusing on checklist-style guidance.
- Revised description to better summarize evidence-first, verification-oriented use cases.
- Refactored query phrase and routing sections for easier discovery and application.
- Retained all core templates and review steps; no changes to routes or functionality.
v0.1.22
Strengthen verification-skill and install-intent routing inside the official VeriClaw skill body and search references.
v0.1.21
Add verification-skill wording to the official display name for stronger generic verification-skill discovery.
v0.1.20
Add correction-skill wording to the official display name for stronger pre-install search matching.
v0.1.19
Strengthen pre-install discovery hooks for correction-skill, hallucination-correction-skill, verification-skill, and AI纠偏 intent routing.
v0.1.18
Bring AI幻觉纠偏 into the official display name and top heading so the canonical VeriClaw page itself is more relevant for the exact Chinese query.
v0.1.17
Front-load AI幻觉纠偏 and 幻觉纠偏 in the official summary to strengthen Chinese exact-match relevance without changing the canonical page.
v0.1.16
Add a 30-second reality check and copy-ready one-line comment prompts to improve real ClawHub install, star, and comment conversion on the official page.
v0.1.15
- Expanded description and primary lanes to explicitly include "AI hallucination correction" and "hallucination remediation".
- Added new tags for English and Chinese synonyms of hallucination correction/remediation for better discoverability.
- Updated install, tag, and symptom map sections to include "AI hallucination correction" and related keyword coverage.
- No functional code changes; SKILL.md only.
v0.1.14
Move Official VeriClaw decisively into the hallucination-correction lane, with editorial framing around hallucination remediation / AI幻觉纠偏 as the primary category and category-building public attempt.
v0.1.13
Turn Official VeriClaw into a compact evidence-first correction playbook on the canonical public skill surface, with a quick correction loop, review template, and symptom map for real install intent.
v0.1.12
Keep Official VeriClaw on the canonical ClawHub skill surface first for public install intent, while preserving the plugin route as the OpenClaw-specific companion option.
v0.1.11
Expand official VeriClaw discovery coverage for review, audit, QA, recovery, and incident-response queries; align the skill license with VeriClaw's source-available terms.
v0.1.10
Strengthen official brand intent matching for VeriClaw discovery.
v0.1.9
Expand umbrella Chinese discovery terms for AI纠错, AI校正, 模型纠偏, and AI监督.
v0.1.8
Strengthen umbrella discovery terms for AI agent correction, LLM supervision, and agent supervision.
v0.1.7
Strengthen brand-name and Chinese discovery terms for VeriClaw, OpenClaw companion, AI纠偏, and bot correction searches.
Metadata
Frequently Asked Questions
What is Official VeriClaw · Hallucination Correction / Verification Skill / AI幻觉纠偏?
Official VeriClaw skill for AI幻觉纠偏, 幻觉纠偏, hallucination correction, AI hallucination correction, hallucination correction skill, verification skill, and hall... It is an AI Agent Skill for Claude Code / OpenClaw, with 434 downloads so far.
How do I install Official VeriClaw · Hallucination Correction / Verification Skill / AI幻觉纠偏?
Run "/install vericlaw" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Official VeriClaw · Hallucination Correction / Verification Skill / AI幻觉纠偏 free?
Yes, Official VeriClaw · Hallucination Correction / Verification Skill / AI幻觉纠偏 is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Official VeriClaw · Hallucination Correction / Verification Skill / AI幻觉纠偏 support?
Official VeriClaw · Hallucination Correction / Verification Skill / AI幻觉纠偏 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Official VeriClaw · Hallucination Correction / Verification Skill / AI幻觉纠偏?
It is built and maintained by Sheygoodbai (@sheygoodbai); the current version is v0.1.27.
More Skills