Security

secureagentics 583 · 1 months ago

Runtime security for AI agents: watches actions AND reasoning traces to catch prompt injection, tool poisoning and out-of-remit behavior — blocking before the action lands. SDKs + Claude Code plugin.

Asymptote-Labs 1.8k · today

Cross-harness memory for coding agents: captures sessions from Claude Code, Cursor, Codex and 20+ harnesses, distills reviewed knowledge over MCP/skills, and forwards telemetry to SIEMs.

eunomia-bpf 743 · yesterday

System-level observability for AI agents via eBPF and TLS tracing: correlates prompts and model calls with the processes, files and network the agent actually touched — no SDK, no proxy.

digitranslab 201 · 8 months ago

Open-source AI security automation (SOAR): visual playbook builder, autonomous triage agents, and 80+ SIEM/EDR/identity/ticketing integrations. Self-hosted, multi-tenant.

cartography-cncf 4.1k · today

CNCF tool that syncs infrastructure assets and relationships (AWS, GCP, Azure, Kubernetes, GitHub, Okta and 30+ more) into a Neo4j graph, then runs security rules and Cypher queries over it.

shuvonsec 5.3k · 6 days ago

Autonomous bug-bounty agent for the terminal — recon, 20 vuln classes, a validation gate and submission-ready HackerOne/Bugcrowd reports. Runs as a Claude Code plugin or standalone on free providers.

vaultry 6 · 6 months ago

Encrypted secrets store for Claude Code: macOS-Keychain-backed vault with MCP server, CLI, commit-safe secret:// .env placeholders and a native input dialog.

calesthio 12.1k · 4 months ago

Self-hosted OSINT terminal: 27 open feeds — satellite fires, flights, radiation, sanctions, markets, conflict data — polled in parallel onto one Jarvis-style dashboard. LLM turns it two-way.

confident-ai 3k · 14 days ago

Open-source red-teaming framework for LLM systems: 50+ vulnerabilities, jailbreak/injection/multi-turn attacks against agents, RAG pipelines and chatbots — plus guardrails. Runs locally.

MichaelS1011 48 · yesterday

Capability-based WASM sandbox for MCP tools and agent-generated code: wasmtime + WASI with fuel, memory, time and I/O caps per run, sub-ms warm execution and signed (JCS/ES256) execution records.

FailproofAI 5.5k · today

Observability plus policy enforcement for agent harnesses: records every run across 12 harnesses (Claude Code, Codex, Cursor, Hermes…) and blocks dangerous tool calls with 39 built-in policies.

openai 157 · today

OpenAI's GitHub Action for CI hardening: bounded egress filtering and runner lockdown — the fence that keeps supply-chain attacks out and your CI-resident coding agent in.

reconurge 9.7k · 12 days ago

Self-hosted OSINT investigation platform — explore entities on a Neo4j-backed visual graph and expand them with 30+ enrichers: DNS/WHOIS/subdomains, breach checks, Maigret, crypto wallets.

NVIDIA 9.6k · yesterday

NVIDIA's LLM vulnerability scanner: nmap-style probing for jailbreaks, prompt injection, data leakage, toxicity and hallucination across dozens of model endpoints.

Atomburstofficial 166 · 24 days ago

Read-only inventory of every AI agent, MCP server, hook, plugin and AI extension on a machine: npx geiger-scan labels what can execute code, hold secrets or reach the filesystem, with evidence.

Giskard-AI 5.9k · 3 days ago

Giskard v3: modular Python evals and red-teaming for agentic systems — scenario-based checks with LLM-as-judge, plus an automatic vulnerability scanner across OWASP LLM Top-10 categories.

NVIDIA-NeMo 7.3k · today

NVIDIA's programmable guardrails for LLM apps: input, output, dialog and retrieval rails defined in Colang, wrapping any model or LangChain runnable.

perplexityai 1.1k · yesterday

Perplexity's endpoint sensor for AI agents: hooks, OTLP logs and on-disk session artifacts normalized into one CEL-matched event model, with opt-in pre-action blocking and offline forensics.

Kritt-ai 2.3k · 3 days ago

Self-hosted platform orchestrating AI agents for vulnerability research: chain focused prompts into reusable workflows, run them in parallel over Codex or Claude Code, dedupe and rank findings.

NVIDIA 16.3k · today

NVIDIA's safe runtime for autonomous agents: kernel-enforced sandboxes with policy on every file, syscall and connection; credentials injected only for approved endpoints.

vxcontrol 30.2k · 6 days ago

Self-hosted autonomous pentesting: multi-agent system in sandboxed Docker with 20+ tools, supervised agent hierarchies, Langfuse observability and a Graphiti knowledge graph. 10+ LLM providers.

luckyPipewrench 929 · today

Agent egress firewall in Go: an inline proxy scanning HTTP, MCP, A2A and WebSocket traffic for secret exfiltration, SSRF and prompt injection, with signed action receipts you verify offline.

zhaoxuya520 42.1k · 23 days ago

Reverse-engineering and pentest skill router for coding agents: routes APK/ELF/JS/PCAP/CTF tasks to the right playbook, bootstraps jadx/Frida/IDA-class toolchains, evolves a knowledge base. CN/EN.

google 1.1k · today

Zero-trust P2P mesh for agents, published by Google: libp2p nodes with portable crypto identities self-discover, and an MCP sidecar lets an agent invoke tools on any node — cloud, local or edge.

NVIDIA 20.2k · today

NVIDIA's security scanner for agent skills: 71 patterns across 17 categories — prompt injection, exfiltration, supply chain, MCP tool poisoning — static plus optional LLM pass, SARIF output.

koala73 88.5k · today

Real-time global situational-awareness dashboard: curated feeds AI-synthesized into briefs over a 3D globe, cross-stream correlation and a country instability index — plus MCP, REST API and CLI.