No red flags found in any of the 11 categories — no credential harvesting, no data exfiltration, no curl-pipe-shell installer. Scanned against the SlowMist agent-security taxonomy, refreshed every 8 hours. Full audit →
by Extraltodeus · LLM Plugin · ★ 228
Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h
🔒 Is J-Wash safe to install? View the security audit →
J-Wash Reshape a model's identity and behavior by editing token directions, then bake those edits into a real checkpoint you can run anywhere. No training, no dataset, no fine-tuning. J-Wash is a local studio (FastAPI + React) for exploring and editing the J-space of any Hugging Face decoder LLM. You chat with a model while a live Jacobian lens shows what each layer is "reading," pin and inspect concepts, then wash the model's identity or behavior with a few token-level rules — turn "I am a large language model" into "I am a large language fish" — and export the result as a standalone model (full checkpoint, modified layers, or LoRA): standard safetensors weights that load anywhere models do. The editing preview runs live in the chat, and the exported checkpoint reproduces it faithfully — the whole point of the project is that what you see is what you ship. Introducing non-expert friendly alignment! What it's built on J-Wash is built on Anthropic's Jacobian lens (the library), a method that reads each layer's contribution to the residual stream through the model's own un-embedding.
| Stars | 228 |
| Forks | 25 |
| Language | Python |
| Category | LLM Plugin |
| License | Apache-2.0 |
| Quality Score | 69.4425312827607/100 |
| Open Issues | 2 |
| Last Updated | 2026-09-06 |
| Created | 2026-07-13 |
| Platforms | python |
| Est. Tokens | ~16k |
Looking for a J-Wash alternative? If you're comparing J-Wash with other llm plugin tools, these 6 projects are the closest alternatives on Agent Skills Hub — ranked by topic overlap, star count, and community traction.
AI-powered universal search for all your personal data, tailored just for you. Goal:The world's first product
AI-powered text compression library for RAG systems and API calls. Reduce token usage by up to 50-60% while pr
Curiso is an infinite canvas for your thoughts
🚀 200倍速！AI时代的下载神器 | Docker/PyPI/HuggingFace/CRAN 全加速 | 并行分片+智能缓存，让下载飞起来
Cutting-edge AI solution for Home Assistant. Multi-LLM provider support to transform your smart home experienc
Ship RAG based LLM web apps in seconds.
Explore other popular llm plugin tools:
J-Wash is Jacobian-Brainwash : A manual alignment tool for large language models built on Anthropic's Jacobian Lens. Results are exportable.. It is categorized as a LLM Plugin with 228 GitHub stars.
J-Wash is primarily written in Python. It covers topics such as abliteration, ai, anthropic.
You can find installation instructions and usage details in the J-Wash GitHub repository at github.com/Extraltodeus/J-Wash. The project has 228 stars and 25 forks, indicating an active community.
J-Wash is released under the Apache-2.0 license, making it free to use and modify according to the license terms.
The top alternatives to J-Wash on Agent Skills Hub include quanta-quest, context-compressor, curiso. Each offers a different approach to the same problem space — compare them side-by-side by stars, quality score, and community activity.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: