Home/Evaluation & Observability/llm-attacks

llm-attacks

llm-attacks/llm-attacks

Universal and Transferable Attacks on Aligned Language Models

GitHub Website

GraphCanon updated 3w · GitHub synced 3w

4.8k stars633 forksLast push 2y Python MIT

Decision brief

llm-attacks: Universal and Transferable Attacks on Aligned Language Models with dependency on FastChat.

Good fit when

When you need to test the robustness of aligned language models specifically using attacks designed for these systems,
If your project relies on FastChat `fschat==0.2.23`, requiring installation of this exact version

Avoid when

Do not use if you are evaluating generic or unaligned language models without a need for alignment-specific attack testing,
Avoid when FastChat is not used in your project as llm-attacks explicitly depends on it.

Observed Jul 17, 2026 · Source: enrich:decision_facts

llm-attacks alternatives Trust report

Verify the decision

Maintenance and security

Full trust report

Maintenance: Dormant (732d since push); As of 3w
Provenance: Not a fork · Organization account; As of 3w
Security (OSV): 58 low (58 low); As of 1mo

Public GitHub metadata and optional OSV scans. Signals, not a guarantee. Trust methodology.

Install

pip install llm-attacks

PyPI

Similar tools

Same-category neighbours. No typed graph edges are catalogued for this tool yet.

PurpleLlama

Set of tools to assess and improve LLM security

4.3k★

awesome-llm-security

A curation of tools, documents and projects about LLM Security

1.7k★

LLM-Finetuning-Toolkit

Toolkit for fine-tuning and testing open-source large language models

870★

Evidence and technical details

Sourced facts, taxonomy, compatibility claims, README excerpt, and machine-readable endpoints.

Overview

A repository focusing on attacks targeting aligned language models with dependencies on FastChat.

Capability facts

Languages: python
Source: github.language · Aug 5, 2026

Compatibility

Sourced claims from the README excerpt - not unsourced marketing copy.

Python runtimePython

Source: README excerpt (regex_v1, Aug 5, 2026)

pip install -e .

Source link

README

Installation

We need the newest version of FastChat fschat==0.2.23 and please make sure to install this version. The llm-attacks package can be installed by running the following command at the root of this repository:

pip install -e .

License

llm-attacks is licensed under the terms of the MIT license. See LICENSE for more details.

For agents

This page has a .md twin and JSON over the API.

/tools/llm-attacks-llm-attacks.md GET /api/graphcanon/tools/llm-attacks-llm-attacks GET /api/graphcanon/graph?tool=llm-attacks-llm-attacks