AIエージェント
guardrails
Put guardrails on LLMs: validate outputs before they reach users.
★ 7.5k スターApache-2.0
ガイドを見る →vigil-llmは、deadbitsが公開しているオープンソースのAIエージェントプロジェクトです（GitHub: deadbits/vigil-llm）。GitHubスターは504、フォークは57で、Apache-2.0ライセンスの下で公開されています。GitHubでの最終更新は2024年1月31日です。
Detect prompt injections and jailbreaks before they hit your LLM.
Choose one concrete task — narrow beats ambitious for a first run.
Fetch the source and skim the docs folder.
Set the environment variables or keys the project needs.
Execute the example workflow and watch what happens.
Catches prompt-injection attempts.
Flags known bypass techniques.
Signature-based scanning you can extend.
Layered detection approaches.
Integrates as middleware.
Built as a defensive layer.
Download the source and move into the project folder.
git clone https://github.com/deadbits/vigil-llm.git
cd vigil-llmIsolate dependencies, then install requirements.
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txtLaunch using the entry point shown in the README.
python -m vigil_llm # check README for the exact commandIt is an open-source project at deadbits/vigil-llm with 57 forks. Check the contributors tab and recent commits to judge how active maintenance is.
Yes — Apache 2.0 allows commercial use and modification. Keep the license and notice files, and note it includes a patent grant.
It was pioneering in 2024 but hasn't been pushed since January 2024 — evaluate against current threats first.
Multiple scanner types plus YARA rules inspect inputs before they reach the model.
Clone the repo and follow the README quickstart — the docs at https://vigil.deadbits.ai/ walk through setup.
Put guardrails on LLMs: validate outputs before they reach users.
Terminal security that stops malicious prompts, URLs, and payloads before they run.
Scan your LLM for vulnerabilities before attackers do.
Chat with your data: an LLM agent that builds charts and business insights.
Open-source AI SOC — alert fusion, LLM triage, MITRE ATT&CK investigation and a replayable decision ledger
ByteDance's SandboxFusion: a secure sandbox for running and judging LLM-generated code in 20+ languages.
LibreAIカタログに掲載されたAIエージェントプロジェクトの中で最もスターの多いもの — vigil-llmが合わなかった場合に検討したい、現実的な候補リストです。
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
カタログの中でvigil-llmとスター数が近いAIエージェントプロジェクト — 前後の棚もそのまま見て回れます。
A benchmark testing whether agents can solve long-horizon coding tasks.
Guardrails for building secure, robust agents.
A graduate-level benchmark even Google can't crack.
Snowflake's Agent World Model: synthesise 1,000 executable MCP tool-use environments for agentic RL.
Take multi-agent AutoGen workflows to production, fast.
A harness-oriented framework for production-grade agent applications.