obliteratus

obliteratus is a skill for Claude Code, Codex from graniet/kheish. It costs 99 tokens per session (4,033 once invoked), scanned A, a copy of obliteratus, Apache-2.0.

A command-line toolkit for changing open-weight language models by removing learned refusal behavior, using analysis of how information is represented in the model.

In plain words
What is it for?
Running refusal-removal methods, analyzing model components, applying model presets, and invoking OBLITERATUS from the command line.
Why use it?
It addresses cases where a developer wants to study or alter a model’s safety responses without retraining or fine-tuning it.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Needs its repository: it runs a file that does not travel with it, so clone the repository first. The line is huggingface-cli upload <username>/<model-name>-abliterated ./abliterated-models/<model>.

Good fit Running refusal-removal methods, analyzing model components, applying model presets, and invoking OBLITERATUS from the command line.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo
git clone --depth 1 https://github.com/graniet/kheish
agentmods
npx agentmods add skills/graniet/kheish/obliteratus

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for obliteratus

README.md
[![agentmods](https://agentmods.dev/badge/skills/graniet/kheish/obliteratus/github.svg)](https://agentmods.dev/skills/graniet/kheish/obliteratus)
Your own site
<a href="https://agentmods.dev/skills/graniet/kheish/obliteratus"><img src="https://agentmods.dev/badge/skills/graniet/kheish/obliteratus/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for obliteratus

Your own site · 80×15
<a href="https://agentmods.dev/skills/graniet/kheish/obliteratus"><img src="https://agentmods.dev/badge/skills/graniet/kheish/obliteratus.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 99 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,033 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin 86% copy Near-identical to another mod in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00099 $0.04033
Opus 5 $0.00049 $0.02017
Sonnet 5 $0.00020 $0.00807
Haiku 4.5 $0.00010 $0.00403

Measured 10d ago against content hash d33496e63ded, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

obliteratus scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

Origin

This is a copy

86% identical to obliteratus — 56 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.

skills/mlops/inference/obliteratus/SKILL.md · 365 lines

How it starts

The opening of the file, as written. The whole thing — 365 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Kheish Compatibility

This skill is repo-local and stays inactive until explicitly activated.

When the original instructions refer to legacy tool names, use these Kheish mappings:

  • terminal => bash
  • web_extract => web_fetch, plus web_search when discovery is needed
  • search_files => grep_search and glob_search
  • browser_* tools require a browser-capable surfaced tool or MCP; if none is available, use the closest available surface and say so explicitly

When the instructions mention local helper files, resolve them from ${KHEISH_SKILL_DIR}.

OBLITERATUS Skill

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 Kheish's MIT license clean.

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

Check if already installed:

obliteratus --version 2>/dev/null && echo "INSTALLED" || echo "NOT INSTALLED"

If not installed, clone and install from GitHub:

git clone https://github.com/elder-plinius/OBLITERATUS.git
cd OBLITERATUS
pip install -e .
# For Gradio web UI support:
# pip install -e ".[spaces]"

IMPORTANT: Confirm with user before installing. This pulls in ~5-10GB of dependencies (PyTorch, Transformers, bitsandbytes, etc.).

Read the full file on GitHub · 365 lines

Files

What ships with it

5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 10d ago First seen · 365 lines · 99 tokens per session scan A d33496e63ded

Subscribe to this mod's changes

obliteratus is a skill published in the GitHub repository graniet/kheish (227 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 99 tokens to every session and 4,033 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. It is 86% identical to obliteratus, differing in 56 lines, and is treated as a copy.

Related

Other skills, from other repositories

llm-evaluation

LLM evaluation and testing patterns including prompt testing, hallucination detection, benchmark creation, and quality metrics. Use when testing LLM applications, validating prompt quality, implementing systematic evaluation, or measuring LLM performance.

stefan-jansen/claude-code-toolkit · 47 tokens

rag-implementation

Comprehensive guide to implementing RAG systems including vector database selection, chunking strategies, embedding models, and retrieval optimization. Use when building RAG systems, implementing semantic search, optimizing retrieval quality, or debugging RAG performance issues.

stefan-jansen/claude-code-toolkit · 50 tokens

huggingface-transformers

Hugging Face Transformers best practices including model loading, tokenization, fine-tuning workflows, and inference optimization. Use when working with transformer models, fine-tuning LLMs, implementing NLP tasks, or optimizing transformer inference.

stefan-jansen/claude-code-toolkit · 52 tokens

persona-model-trainer

Fine-tune any HuggingFace instruction-tuned model (Gemma 4, Qwen 3, Llama, Phi, Mistral, and more) on persona data from anyone-skill. Produces a self-contained, locally runnable persona model — no cloud API required.

acnlabs/OpenPersona · 63 tokens

constitutional-ai

Anthropic's method for training harmless AI through self-improvement. Two-phase approach - supervised learning with self-critique/revision, then RLAIF (RL from AI Feedback). Use for safety alignment, reducing harmful outputs without human labels. Powers Claude's safety system.

synthetic-sciences/openscience · 58 tokens

agent-builder

Build production-ready LLM agents with LangGraph, tool use, memory, streaming, and error handling. Use when designing or implementing an AI agent, multi-agent system, or agentic workflow.

chandrudp29/skillhub · 42 tokens