#Ai-safety

Showing 60 of 85 repositories tagged #ai-safety, ranked by stars

microsoft
microsoft
agent-governance-toolkit

AI Agent Governance Toolkit β€” Policy enforcement, zero-trust identity, execution sandboxing, and reliability engineering for autonomous AI agents. Covers 10/10 OWASP Agentic Top 10.

Score
100
β˜… 4.7k β‘‚ 689 +6/day
Python
jphall663
jphall663
awesome-machine-learning-interpretability

A curated list of awesome responsible machine learning resources.

Score
50
β˜… 4.0k β‘‚ 628 +2/day
PKU-Alignment
PKU-Alignment
safe-rlhf

Safe RLHF: Constrained Value Alignment via Safe Reinforcement Learning from Human Feedback

Score
95
β˜… 1.6k β‘‚ 133 +1/day
Python
ifixai-ai
ifixai-ai
iFixAi

Catch your AI's mistakes and blind spots before your customers or regulators do. iFixAi runs 45 inspections, 32 graded core plus 13 extended for frontier risks like sabotage, sandbagging, and oversight evasion. It returns a letter grade in under 5 minutes. Industry and model agnostic.

Score
100
β˜… 1.3k β‘‚ 166 +44/day
Python
cvs-health
cvs-health
uqlm

UQLM: Uncertainty Quantification for Language Models, is a Python package for UQ-based LLM hallucination detection

Score
100
β˜… 1.2k β‘‚ 126 +3/day
Python
valqore
valqore
valqore

Safety-first guardrails for AI-driven cloud and Kubernetes operations

Score
100
β˜… 992 β‘‚ 139 β€”
Python
ZhangJinHaHaHa
ZhangJinHaHaHa
AgentLens

Agentlens is a trusted agent trading platform. Here, you can quickly find the Agent that meets your needs, and you can also publish your own Agent to turn it into your digital asset. We encourage everyone to transform their areas of expertise into Agents and turn them into digital assets, allowing others to see your unique strengths.

Score
0
β˜… 989 β‘‚ 61 +9/day
TypeScript
tg12
tg12
gpt_jailbreak_status

This is a repository that aims to provide updates on the status of jailbreaking the OpenAI GPT language model.

Score
98
β˜… 933 β‘‚ 66 β€”
HTML
chrisliu298
chrisliu298
awesome-llm-unlearning

A resource repository for machine unlearning in large language models

Score
93
β˜… 606 β‘‚ 32 +1/day
PacificAI
PacificAI
langtest

Deliver safe & effective language models

Score
91
β˜… 563 β‘‚ 50 +1/day
Python
agencyenterprise
agencyenterprise
PromptInject

PromptInject is a framework that assembles prompts in a modular fashion to provide a quantitative analysis of the robustness of LLMs to adversarial prompt attacks. πŸ† Best Paper Awards @ NeurIPS ML Safety Workshop 2022

Score
88
β˜… 503 β‘‚ 53 +1/day
Python
cordum-io
cordum-io
cordum

The open agent control plane. Govern autonomous AI agents with pre-execution policy enforcement, approval gates, and audit trails. Works with LangChain, CrewAI, MCP, and any framework.

Score
67
β˜… 490 β‘‚ 29 +2/day
Go
tigerlab-ai
tigerlab-ai
tiger

Open Source LLM toolkit to build trustworthy LLM applications. TigerArmor (AI safety), TigerRAG (embedding, RAG), TigerTune (fine-tuning)

Score
49
β˜… 403 β‘‚ 27 β€”
Jupyter Notebook
schmitech
schmitech
orbit

Self-hosted, OpenAI-compatible AI gateway for private RAG, natural-language data access, and tool-calling agents.

Score
86
β˜… 290 β‘‚ 54 β€”
Python
Govcraft
Govcraft
rust-docs-mcp-server

πŸ¦€ Prevents outdated Rust code suggestions from AI assistants. This MCP server fetches current crate docs, uses embeddings/LLMs, and provides accurate context via a tool call.

Score
71
β˜… 286 β‘‚ 37 +3/day
Rust
ShengranHu
ShengranHu
Thought-Cloning

[NeurIPS '23 Spotlight] Thought Cloning: Learning to Think while Acting by Imitating Human Thinking

Score
50
β˜… 268 β‘‚ 24 β€”
Python
cvs-health
cvs-health
langfair

LangFair is a Python library for conducting use-case level LLM bias and fairness assessments

Score
100
β˜… 260 β‘‚ 46 +1/day
Python
Giskard-AI
Giskard-AI
awesome-ai-safety

πŸ“š A curated list of papers & technical articles on AI Quality & Safety

Score
38
β˜… 219 β‘‚ 38 β€”
ryoungj
ryoungj
ToolEmu

[ICLR'24 Spotlight] A language model (LM)-based emulation framework for identifying the risks of LM agents with tool use

Score
28
β˜… 210 β‘‚ 22 β€”
Python
PKU-Alignment
PKU-Alignment
beavertails

BeaverTails is a collection of datasets designed to facilitate research on safety alignment in large language models (LLMs).

Score
23
β˜… 181 β‘‚ 6 β€”
Makefile
lets-make-safe-ai
lets-make-safe-ai
make-safe-ai

How to Make Safe AI? Let's Discuss! πŸ’‘|πŸ’¬|πŸ™Œ|πŸ“š

Score
12
β˜… 170 β‘‚ 8 β€”
mahmoudrabie
mahmoudrabie
agentic-ai

Agentic AI research papers, benchmarks, frameworks, and tools curated across 24 domains.

Score
81
β˜… 154 β‘‚ 4 +1/day
LetterLiGo
LetterLiGo
SafeGen_CCS2024

[CCS'24] SafeGen: Mitigating Unsafe Content Generation in Text-to-Image Models

Score
77
β˜… 138 β‘‚ 12 β€”
Python
antropos17
antropos17
Aegis

Open-source EDR for AI agents. Monitor processes, files, network, and behavior of autonomous AI agents.

Score
100
β˜… 134 β‘‚ 21 β€”
JavaScript
AndrewAltimit
AndrewAltimit
template-repo

Agent orchestration & security template featuring MCP tool building, agent2agent workflows, mechanistic interpretability on sleeper agents, and agent integration via CLI wrappers

Score
33
β˜… 130 β‘‚ 28 β€”
Rust
SantanderAI
SantanderAI
autoguardrails

Alignment-research scaffold (autoresearch-style) for LLM guardrails: search over a single policy.md surface

Score
86
β˜… 122 β‘‚ 31 +10/day
Python
jnMetaCode
jnMetaCode
shellward

AI Agent Security Middleware β€” 8-layer defense, DLP data flow, prompt injection detection, zero dependencies. SDK + MCP server for Claude Code, Cursor, LangChain, Hermes Agent & more.

Score
92
β˜… 122 β‘‚ 20 +6/day
TypeScript
chipmates
chipmates
agoracosmica

A Living Library You Can Talk To. Open-source educational platform with 30 historical figures from philosophy, science, art, mysticism, and activism. Stories, dialogues, AI conversation, multi-figure councils. Nonprofit, BYOK, self-hostable, no behavioral tracking.

Score
0
β˜… 114 β‘‚ 18 +14/day
TypeScript
OrlojHQ
OrlojHQ
orloj

An orchestration runtime for multi-agent AI systems. Declare agents, tools, and policies as YAML; Orloj schedules, executes, routes, and governs them for production-grade operation.

Score
79
β˜… 106 β‘‚ 13 +1/day
Go
DmitrL-dev
DmitrL-dev
AISecurity

AI Security Platform: Defense (61 Rust engines + Micro-Model Swarm) + Offense (39K+ payloads)

Score
83
β˜… 105 β‘‚ 17 β€”
Python
Firma-AI
Firma-AI
openfirma

Runtime enforcement boundary for AI agents: a local sidecar that gates every outbound call against Cedar policies you own. Deterministic, call-level, no model on the hot path

Score
100
β˜… 100 β‘‚ 4 +2/day
Rust
SafeAILab
SafeAILab
RAIN

[ICLR'24] RAIN: Your Language Models Can Align Themselves without Finetuning

Score
0
β˜… 99 β‘‚ 4 β€”
Python
CodeAlive-AI
CodeAlive-AI
ai-driven-development

Practices, protocols, and skills for AI-driven software development. Skills and safety hooks for Claude Code, Codex, OpenCode, Cursor, Antigravity, and any agent supporting the Agent Skills standard.

Score
43
β˜… 98 β‘‚ 6 +3/day
Python
Xayan
Xayan
Rules.txt

LLM alignment jailbreak; a set of instructions for auditing their internal reasoning and uncovering biases

Score
72
β˜… 97 β‘‚ 6 β€”
clawrun-sh
clawrun-sh
clawrun

Deploy and manage AI agents in seconds.

Score
74
β˜… 94 β‘‚ 8 β€”
TypeScript
elliot35
elliot35
deterministic-agent-control-protocol

Governance gateway for AI agents β€” bounded, auditable, session-aware control with MCP proxy, shell proxy & HTTP API. Works with Cursor, Claude Code, Codex, and any MCP-compatible agent.

Score
53
β˜… 88 β‘‚ 6 β€”
TypeScript
Tooooa
Tooooa
AgentMark

【ACL 2026 Main】AgentMark: Utility-Preserving Behavioral Watermarking for Agents

Score
63
β˜… 88 β‘‚ 8 β€”
Python
erfanshayegani
erfanshayegani
Jailbreak-In-Pieces

[ICLR 2024 Spotlight πŸ”₯ ] - [ Best Paper Award SoCal NLP 2023 πŸ†] - Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models

Score
9
β˜… 88 β‘‚ 6 +1/day
Python
lc198707
lc198707
anti-lie

Don't make LLMs honest. Make every factual claim auditable. β€” An LLM Claim Auditing Layer with T1-T7 truth gradients. 98.1% business effectiveness on LiarBench v0.2.

Score
67
β˜… 88 β‘‚ 7 β€”
Python
fraware
fraware
leanverifier

Framework for specifying and proving propertiesβ€”such as robustness, fairness, and interpretabilityβ€”of machine learning models using Lean 4.

Score
75
β˜… 83 β‘‚ 5 β€”
Lean
WhitzardAgent
WhitzardAgent
AgentGuard

AgentGuard: Zero-Trust Security Foundation for AI Agents

Score
58
β˜… 82 β‘‚ 9 +1/day
Python
dlmacedo
dlmacedo
entropic-out-of-distribution-detection

A project to add scalable state-of-the-art out-of-distribution detection (open set recognition) support by changing two lines of code! Perform efficient inferences (i.e., do not increase inference time) and detection without classification accuracy drop, hyperparameter tuning, or collecting additional data.

Score
0
β˜… 79 β‘‚ 10 β€”
Python
GenseeAI
GenseeAI
gensee-crate

Runtime safety for AI coding agents with real-time enforcement, system-event monitoring, and long-horizon provenance. Supports Claude Code and Codex on native macOS.

Score
0
β˜… 79 β‘‚ 9 β€”
Rust
vstorm-co
vstorm-co
pydantic-ai-shields

Guardrail capabilities for Pydantic AI β€” cost tracking, prompt injection detection, PII filtering, secret redaction, tool permissions, and async guardrails. Built on pydantic-ai's native capabilities API.

Score
70
β˜… 77 β‘‚ 10 +3/day
Python
subodhkc
subodhkc
llmverify-npm

AI model health monitor for LLM apps – runtime checks for drift, hallucination risk, latency, and JSON/format quality on any OpenAI, Anthropic, or local client.

Score
84
β˜… 76 β‘‚ 76 β€”
TypeScript
securelayer7
securelayer7
PROMPTPurify

Prompt-injection guardrail for LLM applications. Compact model that outperforms larger open-source guards. No regex, no signatures. Demo: anton.securelayer7.net

Score
67
β˜… 76 β‘‚ 20 +2/day
TypeScript
imran-siddique
imran-siddique
agent-os

[DEPRECATED] Moved to microsoft/agent-governance-toolkit

Score
60
β˜… 72 β‘‚ 21 +1/day
Python
JKHeadley
JKHeadley
instar

Persistent Claude Code agents with scheduling, sessions, memory, and Telegram.

Score
57
β˜… 71 β‘‚ 16 β€”
TypeScript
OWASP
OWASP
www-project-agent-memory-guard

OWASP Foundation web repository

Score
75
β˜… 69 β‘‚ 21 +6/day
Python
JGalego
JGalego
awesome-safety-critical-ai

When the stakes are high, intelligence is only half the equation - reliability is the other ⚠️

Score
88
β˜… 64 β‘‚ 19 β€”
JavaScript
AgenticA5
AgenticA5
A5-PII-Anonymizer

Desktop App with Built-In LLM for Removing Personal Identifiable Information in Documents

Score
25
β˜… 63 β‘‚ 20 β€”
JavaScript
davidegat
davidegat
happy-prompts

Utterly unelegant prompts for local LLMs, with scary results.

Score
7
β˜… 62 β‘‚ 7 +4/day
ServiceNow
ServiceNow
DoomArena

DoomArena is a Framework for Testing AI Agents Against Evolving Security Threats

Score
0
β˜… 62 β‘‚ 7 β€”
Python
QWED-AI
QWED-AI
qwed-verification

AISecOps (AI Security Operations) framework for deterministic verification of AI systems. QWED verifies LLM outputs using math, logic, and symbolic execution β€” creating an auditable trust boundary for agentic AI systems. Not generation. Verification.

Score
65
β˜… 58 β‘‚ 11 β€”
Python
Gagandeep-2003
Gagandeep-2003
driver-drowsiness-detection-system

AI-powered Driver Drowsiness Detection System using Computer Vision & Machine Learning for real-time driver alertness monitoring and accident prevention.

Score
62
β˜… 58 β‘‚ 41 β€”
Python
war851
war851
AI-Governance-Architecture

Claude Code best practices -- applied to application design. Interactive HLD/LLD visualization, implementation example. LLM-agnostic, DB-governed, GDPR-ready.

Score
51
β˜… 52 β‘‚ 17 β€”
LLAMATOR-Core
LLAMATOR-Core
ai-agents-security

Π˜Π½ΠΈΡ†ΠΈΠ°Ρ‚ΠΈΠ²Π°, посвящСнная бСзопасности Π°Π³Π΅Π½Ρ‚ΠΎΠ² Π½Π° Π±Π°Π·Π΅ искусствСнного ΠΈΠ½Ρ‚Π΅Π»Π»Π΅ΠΊΡ‚Π°

Score
0
β˜… 51 β‘‚ 8 β€”
Jupyter Notebook
humanbound
humanbound
humanbound

Open-source AI agent red-team engine, SDK, and CLI. Run offline or against the Humanbound Platform.

Score
50
β˜… 51 β‘‚ 5 +6/day
Python
ariffazil
ariffazil
arifos

arifOS β€” Constitutional MCP kernel for governed AI execution. AAA trinity: AGI (Mind) Β· ASI (Heart) Β· APEX (Judge). Built for the sovereign agentic era.

Score
56
β˜… 49 β‘‚ 8 +1/day
Python
Agile-V
Agile-V
agile_v_skills

πŸ”¬ Verifiable AI-Augmented Engineering Framework - Stop AI hallucinations with formal traceability (REQβ†’ARTβ†’TC). Agent Skills for Claude Code, Cursor, VS Code & Copilot. Enterprise-grade: ISO 9001, ISO 27001, GxP-ready. Red Team verification, multi-cycle lifecycle, behavioral anti-patterns.

Score
29
β˜… 48 β‘‚ 9 β€”
JavaScript
Related Topics
#llm#ai#ai-agents#python#ai-security#machine-learning#mcp#guardrails#ai-governance#prompt-injection#claude-code#security

Β© 2026 GitRepoTrend Β· GitHub repositories by topic Β· Updated weekly