Sponsored by Deepsite.site

Tag

#security

245 results found

An open-source security proxy and active firewall for the Model Context Protocol (MCP)

McpVanguard ๐Ÿ›ก๏ธ Titan-Grade AI Firewall for MCP Agents (v1.7.0) MCP (Model Context Protocol) enables AI agents to interact with host-level tools. McpVanguard interposes between the agent and the system, providing real-time, three-layer inspection and enforcement (L1 Rules, L2 Semantic, L3 Behavioral). Transparent integration. Zero-configuration requirements for existing servers. Tests PyPI version License: Apache 2.0 Python 3.11+ Part of the Provnai Open Research Initiative โ€” Building the Immune System for AI. โšก Quickstart pip install mcp-vanguard Local stdio wrap (no network): vanguard start --server "npx @modelcontextprotocol/server-filesystem ." Cloud Security Gateway (SSE, deploy on Railway): export VANGUARD_API_KEY="your-secret-key" vanguard sse --server "npx @modelcontextprotocol/server-filesystem ." Deploy on Railway ๐Ÿ“– Full Railway Deployment Guide ๐Ÿ›ก๏ธ Getting Started (New Users) Bootstrap your security workspace with a single command: # 1. Initialize safe zones and .env template vanguard init # 2. (Optional) Protect your Claude Desktop servers vanguard configure-claude # 3. Launch the visual security dashboard vanguard ui --port 4040 ๐Ÿง  How it works Every time an AI agent calls a tool (e.g. read_file, run_command), McpVanguard inspects the request across three layers before it reaches the underlying server: Layer What it checks Latency L1 โ€” Safe Zones & Rules Kernel-level isolation (openat2 / Windows canonicalization) and 50+ deterministic signatures ~16ms L2 โ€” Semantic LLM-based intent scoring via OpenAI, DeepSeek, Groq or Ollama Async L3 โ€” Behavioral Shannon Entropy ($H(X)$) scouter and sliding-window anomaly detection Stateful Performance Note: The 16ms overhead is measured at peak concurrent load. In standard operation, the latency is well under 2msโ€”negligible relative to typical LLM inference times. If a request is blocked, the agent receives a standard JSON-RPC error response. The underlying server never sees it. Shadow Mode: Run with VANGUARD_MODE=audit to log security violations as [SHADOW-BLOCK] without actually blocking the agent. Perfect for assessing risk in existing production workflows. ๐Ÿ›ก๏ธ What gets blocked Sandbox Escapes: TOCTOU symlink attacks, Windows 8.3 shortnames (PROGRA~1), DOS device namespaces Data Exfiltration: High-entropy payloads (H > 7.5 cryptographic keys) and velocity-based secret scraping Filesystem attacks: Path traversal (../../etc/passwd), null bytes, restricted paths (~/.ssh), Unicode homograph evasion Command injection: Pipe-to-shell, reverse shells, command chaining via ; && \n, expansion bypasses SSRF & Metadata Protection: Blocks access to cloud metadata endpoints (AWS/GCP/Azure) and hex/octal encoded IPs. Jailbreak Detection: Actively identifies prompt injection patterns and instruction-ignore sequences. Continuous Monitoring: Visualize all of the above in real-time with the built-in Security Dashboard. ๐Ÿ“Š Security Dashboard Launch the visual monitor to see your agent's activity and security status in real-time. vanguard ui --port 4040 The dashboard provides a low-latency, HTMX-powered feed of: Real-time Blocks: Instantly see which rule or layer triggered a rejection. Entropy Scores: Pulse-check the H ( X ) levels of your agent's data streams. Audit History: Contextual log fragments for rapid incident response. VEX Protocol โ€” Deterministic Audit Log When McpVanguard blocks an attack, it creates an OPA/Cerbos-compatible Secure Tool Manifest detailing the Principal, Action, Resource, and environmental snapshot. This manifest is then sent as a cryptographically-signed report to the VEX Protocol. VEX anchors that report to the Bitcoin blockchain via the CHORA Gate. This means an auditor can independently verify exactly what was blocked, the entropy score, and why โ€” without relying on your local logs. export VANGUARD_VEX_URL="https://api.vexprotocol.com" export VANGUARD_VEX_KEY="your-agent-jwt" export VANGUARD_AUDIT_FORMAT="json" # Optional: Route JSON logs directly into SIEM (ELK, Splunk) vanguard sse --server "..." --behavioral Architecture โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ” AI Agent โ”‚ McpVanguard Proxy โ”‚ (Claude, GPT) โ”‚ โ”‚ โ”‚ โ”‚ โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ” โ”‚ โ”‚ JSON-RPC โ”‚ โ”‚ L1 โ€” Rules Engine โ”‚ โ”‚ โ”‚โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ–ถโ”‚ โ”‚ 50+ YAML signatures (path, cmd, net...) โ”‚ โ”‚ โ”‚ (stdio/SSE) โ”‚ โ”‚ BLOCK on match โ†’ error back to agent โ”‚ โ”‚ โ”‚ โ”‚ โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜ โ”‚ โ”‚ โ”‚ โ”‚ pass โ”‚ โ”‚ โ”‚ โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ–ผโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ” โ”‚ โ”‚ โ”‚ โ”‚ L2 โ€” Semantic Scorer (optional) โ”‚ โ”‚ โ”‚ โ”‚ โ”‚ OpenAI / MiniMax / Ollama scoring 0.0โ†’1.0โ”‚ โ”‚ โ”‚ โ”‚ โ”‚ Async โ€” never blocks the proxy loop โ”‚ โ”‚ โ”‚ โ”‚ โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜ โ”‚ โ”‚ โ”‚ โ”‚ pass โ”‚ โ”‚ โ”‚ โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ–ผโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ” โ”‚ โ”‚ โ”‚ โ”‚ L3 โ€” Behavioral Analysis (optional) โ”‚ โ”‚ โ”‚ โ”‚ โ”‚ Sliding window: scraping, enumeration โ”‚ โ”‚ โ”‚ โ”‚ โ”‚ In-memory or Redis (multi-instance) โ”‚ โ”‚ โ”‚ โ”‚ โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜ โ”‚ โ”‚ โ”‚ โ”‚ โ”‚ โ”‚โ—€โ”€โ”€ BLOCK โ”€โ”€โ”€โ”€โ”€โ”‚โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ค (any layer) โ”‚ โ”‚ (JSON-RPC โ”‚ โ”‚ ALLOW โ”‚ โ”‚ error) โ”‚ โ–ผ โ”‚ โ”‚ โ”‚ MCP Server Process โ”‚ โ”‚ โ”‚ (filesystem, shell, APIs...) โ”‚ โ”‚ โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜ โ”‚ โ”‚ โ”‚โ—€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€ response โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜ โ”‚ โ”‚ (on BLOCK) โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ–ถ VEX API โ”€โ”€โ–ถ CHORA Gate โ”€โ”€โ–ถ Bitcoin Anchor (async, fire-and-forget audit receipt) L2 Semantic Backend Options The Layer 2 semantic scorer supports a Universal Provider Architecture. Set the corresponding API keys to activate a backend โ€” the first available key wins (priority: Custom > OpenAI > MiniMax > Ollama): Backend Env Vars Notes Universal Custom (DeepSeek, Groq, Mistral, vLLM) VANGUARD_SEMANTIC_CUSTOM_KEY, VANGUARD_SEMANTIC_CUSTOM_MODEL, VANGUARD_SEMANTIC_CUSTOM_URL Fast, cheap inference. Examples: Groq: https://api.groq.com/openai/v1 DeepSeek: https://api.deepseek.com/v1 OpenAI VANGUARD_OPENAI_API_KEY, VANGUARD_OPENAI_MODEL Default model: gpt-4o-mini MiniMax VANGUARD_MINIMAX_API_KEY, VANGUARD_MINIMAX_MODEL, VANGUARD_MINIMAX_BASE_URL Default model: MiniMax-M2.5 Ollama (local) VANGUARD_OLLAMA_URL, VANGUARD_OLLAMA_MODEL Default model: phi4-mini. No API key required # Example: Use Groq for ultra-fast L2 scoring export VANGUARD_SEMANTIC_ENABLED=true export VANGUARD_SEMANTIC_CUSTOM_KEY="your-groq-key" export VANGUARD_SEMANTIC_CUSTOM_MODEL="llama3-8b-8192" export VANGUARD_SEMANTIC_CUSTOM_URL="https://api.groq.com/openai/v1" vanguard start --server "npx @modelcontextprotocol/server-filesystem ."

Turbopentest

TurboPentest is an agentic AI penetration testing platform built by IntegSec that makes professional-grade security assessments accessible to every organization. Instead of waiting weeks for a manual pentest engagement or relying on a single scanning tool, TurboPentest deploys up to 20 autonomous AI agents that orchestrate 15 professional security tools - including Nmap, OWASP ZAP, Nuclei, Nikto2, OpenVAS, TestSSL, Subfinder, HTTPX, FFUF, Wafw00f, Gitleaks, Semgrep, Trivy, IntegSec PentestTools, and Paladin AI - in a single automated workflow. At the core is Paladin AI, an autonomous pentesting agent powered by Claude Sonnet 4.6. While Phase 1 tools gather reconnaissance data, Paladin is where the actual penetration test happens - specialist AI agents validate exploits, discover multi-step attack chains, generate proof-of-concept demonstrations, and assess business impact. This is what makes TurboPentest an agentic pentest rather than just a scanner with AI features. The result is a comprehensive security assessment that covers the OWASP Top 10, network-level vulnerabilities, TLS/SSL misconfigurations, exposed secrets, hidden endpoints, subdomain enumeration, and more. Every scan generates five professional deliverables: a detailed PDF report with severity-ranked findings and remediation guidance, a blockchain-anchored security attestation letter (Base L2) suitable for auditors, customers, and compliance frameworks like SOC 2 and ISO 27001, an attack surface map visualizing your external exposure, a STRIDE-based threat model outlining risk scenarios with prioritized manual testing recommendations, and retest commands (Docker-based) to verify remediation. Scans support both black box (external-only) and white box (with GitHub integration for source-level analysis via Gitleaks, Semgrep SAST, and Trivy SCA) modes. TurboPentest integrates directly into CI/CD pipelines through its GitHub Action, enabling security testing on every pull request or deployment. Results are typically delivered in up to 4 hours. Pricing starts at $49 per scan (Recon tier). Four tiers are available: Recon ($49, 1 agent, 30 min), Standard ($99, 4 agents, 1 hour), Deep ($299, 10 agents, 2 hours), and Blitz ($699, 20 agents, 4 hours). Volume discounts of 10-30% are available for 10+ scans, and subscription plans offer additional savings. TurboPentest is hosted on Microsoft Azure. Reports are built to meet the documentation requirements of SOC 2, ISO 27001, PCI DSS, HIPAA, and CMMC. For more information, visit https://turbopentest.com or contact support@integsec.com.

็ผ”้›ถ็ง‘ๆŠ€mcp

็ผ”้›ถๆณ•ๅˆ™MCPๆ˜ฏไธ“ไธบๅฎกๆ ธๅฎ‰ๅ…จ้ข†ๅŸŸๆ‰“้€ ็š„AIๆ™บ่ƒฝ้˜ฒๅพกไธญๆžข๏ผŒ่‡ดๅŠ›ไบŽ้€š่ฟ‡Agentic AI ๆŠ€ๆœฏๅ‡ๅฐ‘ไบบๅŠ›ๆˆๆœฌ๏ผŒ้ซ˜ๆ•ˆ้ซ˜็ฒพๅบฆไธบ็”จๆˆทๆไพ›ๅˆ†้’Ÿ็บงๆŽฅๅ…ฅ็š„ๅ†…ๅฎน้ฃŽๆŽง่งฃๅ†ณๆ–นๆกˆ๏ผŒ็ ด่งฃๅฎ‰ๅ…จๅจ่ƒ๏ผŒๆไพ›ไปŽ้ฃŽ้™ฉๆ„Ÿ็ŸฅๅˆฐไธปๅŠจๆ‹ฆๆˆช็ญ–็•ฅๆ‰ง่กŒ็š„ๅ…จ้“พ่ทฏ้—ญ็ŽฏไธŽไธ€ไฝ“ๅŒ–่งฃๅ†ณๆ–นๆกˆใ€‚ ๆ ธๅฟƒ็‰น็‚น ไฝฟ็”จ็ฎ€ๅ•ๆŽฅๅ…ฅไพฟๆท๏ผšๆ— ้œ€้ƒจ็ฝฒๆœฌๅœฐๆœๅŠก๏ผŒ้€š่ฟ‡็ฎ€ๅ•้…็ฝฎๅณๅฏไฝฟ็”จ๏ผ› ่ฟ็ปดๆˆๆœฌ่พƒไฝŽ๏ผš็”จๆˆทๆ— ้œ€ๅ…ณๅฟƒๆœๅŠกๅ™จ็ปดๆŠคใ€่ต„ๆบๆ‰ฉๅฎน็ญ‰ๅบ•ๅฑ‚่ฟ็ปด้—ฎ้ข˜ใ€‚ ๅ่ฎฎๅ…ผๅฎน๏ผšๅคšๆจกๆ€ๆ”ฏๆŒ๏ผŒ้€‚้…ไธๅŒๅฎขๆˆทไธๅŒไธšๅŠกไธๅŒๅ†…ๅฎนๅฎกๆ ธๅœบๆ™ฏ็š„ๆŠ€ๆœฏๆŽฅๅ…ฅ้œ€ๆฑ‚ใ€‚ ๆ˜“้›†ๆˆ๏ผšๆ ‡ๅ‡† MCP ๅ่ฎฎ๏ผŒไธปๆต AI/Agent/workflow ็š†ๅฏ็”จ ไธ‰ใ€ๆ ธๅฟƒๅŠŸ่ƒฝ็Ÿฉ้˜ต 1ใ€ๅ…จ้ฃŽ้™ฉๅŸŸ่ฆ†็›– ไธ€ๆฌกๆ€ง่ฏ†ๅˆซๆ‰€ๆœ‰ๅ†…ๅฎน้ฃŽ้™ฉ็ฑปๅž‹๏ผŒๆ— ้œ€ๅคš็ณป็ปŸๅˆ‡ๆขใ€‚ๆœ€้ซ˜ๆ”ฏๆŒ30็ง้ฃŽ้™ฉๅŸŸ๏ผŒๆ”ฏๆŒ 350+็ป†ๅˆ†ๆ ‡็ญพ๏ผŒไธ€ๆฌก่ฐƒ็”จๅฎŒๆˆๅ…จๅŸŸๆ‰ซๆ๏ผŒๆ— ้œ€้‡ๅคๅฏนๆŽฅไธๅŒๅฎกๆ ธๆŽฅๅฃ 2ใ€ๆž้€Ÿๅ“ๅบ”่ƒฝๅŠ› ๅฐ†ๅฎกๆ ธๆ•ˆ็އๆๅ‡่‡ณไบบๅทฅ็š„ 50ๅ€๏ผŒๅนณๅ‡ๅฎกๆ ธๆ—ถ้—ด20็ง’๏ผ›่ง„ๅˆ™ๅ˜ๅŠจๆ—ถ 30ๅˆ†้’ŸๅฎŒๆˆ็ญ–็•ฅ่ฐƒๆ•ด๏ผ›ๆ–ฐ้ฃŽ้™ฉๅœบๆ™ฏ 5ๅคฉๅฟซ้€ŸๆŽฅๅ…ฅ้€‚้… 3ใ€ๆ™บ่ƒฝ่ฏ†ๅˆซๅผ•ๆ“Ž ้€š่ฟ‡ๆทฑๅบฆ่ฏญไน‰ๅˆ†ๆžๅบ”ๅฏนๅคๆ‚้ฃŽ้™ฉๅœบๆ™ฏ๏ผŒ้€š่ฟ‡ไธŠไธ‹ๆ–‡ๅ…ณ่”้‡ๅปบ๏ผŒ่งฃๆž่ฏ†ๅˆซๅ˜ไฝ“่กจ่พพๅ’Œ้š่—ๅซไน‰๏ผ›ๅŒๆ—ถ้€š่ฟ‡ๅŠจๆ€็ญ–็•ฅไผ˜ๅŒ–๏ผŒๆ นๆฎๅฎžๆ—ถ้ฃŽ้™ฉๆ•ฐๆฎ่‡ชๅŠจ่ฐƒๆ•ดๅฎกๆ ธ้˜ˆๅ€ผ๏ผŒ่‡ชๅŠจๅญฆไน ๆ–ฐๅž‹่ฟ่ง„ๆจกๅผใ€‚ 4ใ€ๅคšๆจกๆ€่žๅˆๅฎกๆŸฅ ๅฏ็ปŸไธ€ๅค„็†ๆ–‡ๆœฌใ€ๅ›พๅƒๅ†…ๅฎนใ€‚้€š่ฟ‡ๅ›พๆ–‡ๆททๅˆๅฎกๆ ธๆ—ขๅณๅฏๅฎž็ŽฐAIGCๅ†…ๅฎนไธ“้กนๆฒป็†๏ผŒๆ‹ฆๆˆช่ฟ่ง„็”Ÿๅ›พๆ็คบ่ฏๅŠ่พ“ๅ‡บใ€‚ 5ใ€ไผไธš็บง้ฃŽ้™ฉ็ฎก็†ๅทฅๅ…ท ๅฎขๆˆทๅฏๆ— ๅŸน่ฎญๆˆๆœฌ็›ดๆŽฅไฝฟ็”จ๏ผ›ๅŒๆ—ถๅ…ทๅค‡็ญ–็•ฅ้…็ฝฎไธญๅฟƒ๏ผŒไปฅ่‡ช็„ถ่ฏญ่จ€ๅฎšไน‰่ง„ๅˆ™ไธบๅŸบ็ก€่ฎพ็ฝฎๅˆ†็บงๅฎกๆ ธๆœบๅˆถ๏ผŒ็ฒพๅ‡†ๅฎกๆ ธ