nifi-CVE-2026-39816-poc

Apache NiFi CVE-2026-39816 POC CVE-2026-39816 の概念実証デモ。Apache NiFi 2.8.0 における EXECUTECODE 権限バイパスで、コード実行権限を意図的に拒否されたフローデザイナーが、 グラフバンドルの ExecuteGraphQuery プロセッサを介して NiFi JVM 内で 任意の Groovy を実行できるようになります。 NiFi 2.9.0 で修正済み ZeroPath によって発見 -- 完全な技術解説はこちらで公開されています。 脆弱性 TinkerpopClientService における...

GHSA-2XV2-W8CQ-5GXW PraisonAI: CodeAgent Executes LLM-Generated Code Without Sandboxing and Leaks All Environment Secrets

Summary CodeAgent.executepython executes LLM-generated Python code in a subprocess with the complete parent-process environment os.environ.copy, zero AST validation, zero import restrictions, and no sandbox enforcement — even when CodeConfigsandbox=True is explicitly set. This allows an attacker...

PraisonAI: CodeAgent Executes LLM-Generated Code Without Sandboxing and Leaks All Environment Secrets

Summary CodeAgent.executepython executes LLM-generated Python code in a subprocess with the complete parent-process environment os.environ.copy, zero AST validation, zero import restrictions, and no sandbox enforcement — even when CodeConfigsandbox=True is explicitly set. This allows an attacker...

PraisonAI: Human-in-the-loop tool approval is cached by tool name and silently reused for all subsequent calls with arbitrary arguments

Summary PraisonAI gates dangerous tools file writes, deletes, shell/code execution behind an interactive approval prompt. The first approval of a tool is cached for the remainder of the run and silently reused for all later invocations of that tool with arbitrary, unreviewed arguments. Root cause...

GHSA-29R9-67VG-QJ56 PraisonAI: Human-in-the-loop tool approval is cached by tool name and silently reused for all subsequent calls with arbitrary arguments

Summary PraisonAI gates dangerous tools file writes, deletes, shell/code execution behind an interactive approval prompt. The first approval of a tool is cached for the remainder of the run and silently reused for all later invocations of that tool with arbitrary, unreviewed arguments. Root cause...

bouvet

Bouvet AI 에이전트를 위한 격리된 코드 실행 샌드박스 소개 • 작동 방식 • MCP 도구 • 문서 Bouvet란? Bouvet boo-veh는 AI 에이전트가 코드를 실행할 수 있는 안전하고 격리된 샌드박스를 생성하는 MCP 서버입니다. AI 에이전트가 Python, Node.js 또는 셸 명령을 실행해야 할 때, Bouvet는 약 200ms 만에 경량 마이크로VM을 가동합니다. 코드는 완전히 격리된 환경별도의 커널, 파일 시스템, 네트워크에서 실행된 후 샌드박스가 파괴됩니다. 아무것도 유지되지 않고, 아무것도 새지...

agent-governance-toolkit

🌍 English | 日本語 | 简体中文 | 한국어 Agent Governance Toolkit 에이전트를 프로덕션에 안심하고 배포하세요 🚀빠른 시작 · 📋 명세 · 📦 PyPI · 📝 변경 이력 !IMPORTANT 공개 프리뷰Public Preview — 프로덕션 품질의 Microsoft 서명 릴리즈입니다. GA 이전에 주요 변경이 발생할 수 있습니다. 자율형 AI 에이전트를 위한 정책 적용, 신원증명, 샌드박싱, SRE. pip install 하나로 어떤 프레임워크에서도 사용 가능합니다. 문제 상황 여러분의 AI 에이전트는...

CVE-2025-3248-POC

Langflow RCE エクスプロイト - CVE-2025-3248 ⚠️ 免責事項 このコードは教育および研究目的 のみを目的としています。明示的な許可なくシステムで無断使用することは、法律違反となる可能性があります。 作者は、このツールの誤用や悪意のある活動について一切の責任を負いません 。 脆弱性情報 CVE-2025-3248 - 重要 CVSS スコア: 9.8 | EPSS スコア: 92.57% | 公開日: 2025年5月22日 影響を受けるバージョン 脆弱なバージョン: v1.3.0 以下のすべてのバージョン 修正済み: v1.3.0 以降のバージョン 攻撃ベクトル...

cheatengine-mcp-bridge

데모 Cheat Engine MCP 브리지 수십억 달러 규모 AI 데이터센터가 프로그램 메모리를 분석하게 하세요. 모드, 트레이너, 보안 감사, 게임 봇을 만들고, 리버스 엔지니어링을 가속화하거나, 어떤 프로그램과 게임이든 훨씬 짧은 시간에 처리할 수 있습니다. !NOTE 모든 분들의 별에 감사드립니다. 정말 감사합니다! 특히 모든 기여자분들께 큰 감사를 드립니다!! @libangli218, @lauralex, @iamtyroon, @HachiroSan, @Attacktive 문제 상황 수 기가바이트의 메모리를 바라보고 있습니...

CVE-2026-53710

MCP Context Forge is an AI gateway, registry, and proxy for MCP, A2A, REST, and gRPC APIs. Prior to 1.0.2, the pythonsandboxserver in mcp-servers/python/pythonsandboxserver/src/pythonsandboxserver/serverfastmcp.py exposes raw getattr through safebuiltins, omits a required getattr guard, and relie...

CVE-2026-53710

MCP Context Forge is an AI gateway, registry, and proxy for MCP, A2A, REST, and gRPC APIs. Prior to 1.0.2, the pythonsandboxserver in mcp-servers/python/pythonsandboxserver/src/pythonsandboxserver/serverfastmcp.py exposes raw getattr through safebuiltins, omits a required getattr guard, and relie...

CVE-2026-53710 MCP Context Forge: RestrictedPython sandbox bypass via getattr builtin in python_sandbox_server

MCP Context Forge is an AI gateway, registry, and proxy for MCP, A2A, REST, and gRPC APIs. Prior to 1.0.2, the pythonsandboxserver in mcp-servers/python/pythonsandboxserver/src/pythonsandboxserver/serverfastmcp.py exposes raw getattr through safebuiltins, omits a required getattr guard, and relie...

CVE-2026-53710 MCP Context Forge: RestrictedPython sandbox bypass via getattr builtin in python_sandbox_server

MCP Context Forge is an AI gateway, registry, and proxy for MCP, A2A, REST, and gRPC APIs. Prior to 1.0.2, the pythonsandboxserver in mcp-servers/python/pythonsandboxserver/src/pythonsandboxserver/serverfastmcp.py exposes raw getattr through safebuiltins, omits a required getattr guard, and relie...

EUVD-2026-78933

MCP Context Forge is an AI gateway, registry, and proxy for MCP, A2A, REST, and gRPC APIs. Prior to 1.0.2, the pythonsandboxserver in mcp-servers/python/pythonsandboxserver/src/pythonsandboxserver/serverfastmcp.py exposes raw getattr through safebuiltins, omits a required getattr guard, and relie...

CVE-2026-53710

MCP Context Forge (IBM) is an AI gateway and proxy for MCP, A2A, REST, and gRPC APIs. The vulnerability affects the python_sandbox_server subproject (prior to 1.0.2 ), where the RestrictedPython sandbox can be bypassed through three compounding weaknesses: raw getattr exposed in safe_builtins, ru...

CVE-2026-53710 MCP Context Forge: RestrictedPython sandbox bypass via getattr builtin in python_sandbox_server

MCP Context Forge is an AI gateway, registry, and proxy for MCP, A2A, REST, and gRPC APIs. Prior to 1.0.2, the pythonsandboxserver in mcp-servers/python/pythonsandboxserver/src/pythonsandboxserver/serverfastmcp.py exposes raw getattr through safebuiltins, omits a required getattr guard, and relie...

CVE-2026-53710: Improper Control of Generation of Code

MCP Context Forge is an AI gateway, registry, and proxy for MCP, A2A, REST, and gRPC APIs. Prior to 1.0.2, the pythonsandboxserver in mcp-servers/python/pythonsandboxserver/src/pythonsandboxserver/serverfastmcp.py exposes raw getattr through safebuiltins, omits a required getattr guard, and relie...

CVE-2026-57120

PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.59, executecode sandbox mode permits runtime assembly of blocklisted dunder names and allows str.format or str.formatmap to resolve dotted fields through C-level attribute access that bypasses safegetattr. This exposes class,...

CVE-2026-57120 PraisonAI: execute_code sandbox bypass: str.format C-level attribute access reads every blocklisted dunder

PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.59, executecode sandbox mode permits runtime assembly of blocklisted dunder names and allows str.format or str.formatmap to resolve dotted fields through C-level attribute access that bypasses safegetattr. This exposes class,...

CVE-2026-57120

PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.59, executecode sandbox mode permits runtime assembly of blocklisted dunder names and allows str.format or str.formatmap to resolve dotted fields through C-level attribute access that bypasses safegetattr. This exposes class,...