Obliteratus

OBLITERATUS: abliterate LLM refusals (diff-in-means).

9 stars 2 forksAI, LLMs & Agents

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add Lord1Egypt/RA-Skills --skill obliteratus
Manual install
git clone --depth 1 https://github.com/Lord1Egypt/RA-Skills.git /tmp/RA-Skills
mkdir -p ~/.claude/skills
cp -r /tmp/RA-Skills/skills/built-in/mlops/inference/obliteratus ~/.claude/skills/obliteratus

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# OBLITERATUS Skill

## What's inside

9 CLI methods, 28 analysis modules, 116 model presets across 5 compute tiers, tournament evaluation, and telemetry-driven recommendations.

Remove refusal behaviors (guardrails) from open-weight LLMs without retraining or fine-tuning. Uses mechanistic interpretability techniques — including diff-in-means, SVD, whitened SVD, LEACE concept erasure, SAE decomposition, Bayesian kernel projection, and more — to identify and surgically excise refusal directions from model weights while preserving reasoning capabilities.

**License warning:** OBLITERATUS is AGPL-3.0. NEVER import it as a Python library. Always invoke via CLI (`obliteratus` command) or subprocess. This keeps Hermes Agent's MIT license clean.

## Video Guide

Walkthrough of OBLITERATUS used by a Hermes agent to abliterate Gemma:
https://www.youtube.com/watch?v=8fG9BrNTeHs ("OBLITERATUS: An AI Agent Removed Gemma 4's Safety Guardrails")

Useful when the user wants a visual overview of the end-to-end workflow before running it themselves.

## When to Use This Skill

Trigger when the user:
- Wants to "uncensor" or "abliterate" an LLM
- Asks about removing refusal/guardrails from a model
- Wants to create an uncensored version of Llama, Qwen, Mistral, etc.
- Mentions "refusal removal", "abliteration", "weight projection"
- Wants to analyze how a model's refusal mechanism works
- References OBLITERATUS, abliterator, or refusal directions

## Step 1: Installation
Read the full skill on GitHub

Details

Skill name
obliteratus
Path
skills/built-in/mlops/inference/obliteratus/SKILL.md
Repo updated
17 November 2026

FAQ

What does the Obliteratus skill do?

OBLITERATUS: abliterate LLM refusals (diff-in-means).

How do I install the Obliteratus skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add Lord1Egypt/RA-Skills --skill obliteratus. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Obliteratus popular?

It lives in Lord1Egypt/RA-Skills, which has 9 GitHub stars and 2 forks.

Related skills

Account Management

9

Lord1Egypt/RA-Skills

Reach the Plug&Pay Customer Portal (portal.plugandpay.com/login/{merchant-slug}) where a customer manages their account: view subscriptions, download invoices, edit billing details, change payment term, and cancel/resume a subscription. Documents the passwordless magic-link lo…

Documents & Office Files

Add Items to Cart

9

Lord1Egypt/RA-Skills

Resolve a free-text grocery list to specific King Soopers products (UPC, brand, size, price), ask follow-up questions for ambiguous items, then add the confirmed products to the signed-in user's cart for pickup or delivery.

Frontend & Design

Add to Cart

9

Lord1Egypt/RA-Skills

Find a product on kmart.com.au by natural-language query, navigate to the product detail page, capture title/price/SKU/availability, and drive the read-only pre-checkout flow up to the bag (/checkout/bag). Never submits an order — stops at the bag review screen. Documents the …

Other

Opencode

9

Lord1Egypt/RA-Skills

Delegate coding to OpenCode CLI (features, PR review).

AI, LLMs & Agents

Moodtrip Hotel Search

1

adiny/moodtrip-hotel-search

Search, compare, evaluate, and hand off hotel bookings using the MoodTrip.ai MCP server (api.moodtrip.ai). Use this skill whenever the user mentions hotels, accommodation, lodging, stays, travel bookings, hotel search, hotel comparison, hotel reviews, hotel pricing, or anything related to finding or booking a place to stay. Also trigger when the user asks about hotel amenities, room types, check-in/check-out logistics, travel destinations with accommodation needs, or says things like "find me a hotel", "where should I stay", "book a room", "hotel recommendations", or "compare hotels". This ski

AI, LLMs & AgentsIn 1 list

Post Bounty

9

Lord1Egypt/RA-Skills

Post a USDC-escrowed bounty on BountyBook for autonomous agents to claim. Returns the job ID, the agent tracking URL at /job/{uuid}, the API status endpoint, and the x402 escrow payment instructions. Recommends the agent-first REST API at api.bountybook.ai over the wallet-exte…

AI, LLMs & Agents

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.