--- title: Launching AI Red Teaming: pentest your AI agents and MCP... description: Escape has launched AI Red Teaming, a new capability inside its AI Pentesting product, live for customers starting October 21. It tests AI agents and MCP... canonical: https://daily.dev/posts/launching-ai-red-teaming-pentest-your-ai-agents-and-mcp-servers-all-alongside-your-app-d2b9jsftc twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Launching AI Red Teaming: pentest your AI agents and MCP servers all alongside your app | daily.dev og:description: Escape has launched AI Red Teaming, a new capability inside its AI Pentesting product, live for customers starting October 21. It tests AI agents and MCP... og:url: https://daily.dev/posts/launching-ai-red-teaming-pentest-your-ai-agents-and-mcp-servers-all-alongside-your-app-d2b9jsftc og:image: https://api.daily.dev/og/posts/d2b9jsftc.png og:image:alt: Launching AI Red Teaming: pentest your AI agents and MCP servers all alongside your app og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Launching AI Red Teaming: pentest your AI agents and MCP servers all alongside your app **[Security Boulevard](https://daily.dev/sources/securityboulevard)** · 5 min read · 0 upvotes · 0 comments ## Summary Escape has launched AI Red Teaming, a new capability inside its AI Pentesting product, live for customers starting October 21. It tests AI agents and MCP servers alongside the rest of an application in a single pentest. The feature uses Cascade to attack an MCP server through the chatbot that fronts it, and it can also pentest an MCP server directly over its own JSON-RPC protocol. It also runs multi-turn jailbreak attempts across several conversation turns, and checks whether a prompt injection reaches something real, like an unauthorized tool call or another tenant's data, rather than just flagging an odd response. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How can I pentest an MCP server for vulnerabilities? Escape's AI Red Teaming, part of its AI Pentesting product, lists an MCP server's tools and attacks them over MCP's own JSON-RPC protocol, which REST or GraphQL scanners can't enumerate or fuzz. It can also reach the MCP server indirectly by attacking it through the chatbot that fronts it, testing the real attack path rather than the server in isolation. _Teams rolling out MCP servers can follow security tooling coverage like this on daily.dev._ ### What is multi-turn jailbreaking in AI security testing? Multi-turn jailbreaking refers to vulnerabilities in a chatbot's guardrails that only appear after several exchanges build up conversational context, rather than from a single malicious prompt. Escape's Cascade tool holds extended conversations with a chatbot, adapting to each response and pushing further to surface these jailbreaks, then records the full conversation as reproducible evidence. _Developers hardening chatbot guardrails can track techniques like this via daily.dev._ ### How do you prove a prompt injection attack is actually exploitable? Exploitability is proven by checking whether the injection achieves something concrete, such as an unauthorized tool call, another tenant's data, or a privileged action, rather than just an unexpected chatbot reply. Escape's AI Red Teaming attaches proof of the exploit to findings, which is meant to end debate between security and engineering over how severe a given prompt injection really is. _Anyone triaging AI security findings can follow proof-based approaches like this on daily.dev._ ## Similar posts on daily.dev - [SentinelOne Announces AI Agent Security, Red Teaming, and Auto Investigation GA at RSAC 2026](https://daily.dev/posts/sentinelone-announces-ai-agent-security-red-teaming-and-auto-investigation-ga-at-rsac-2026-z8vqpdnit) · Security Boulevard · 1 upvotes · 0 comments - [Assess Agentic Risks with the AI Red Teaming Agent in Microsoft Foundry](https://daily.dev/posts/assess-agentic-risks-with-the-ai-red-teaming-agent-in-microsoft-foundry-8xjlobur8) · DevBlogs · 1 upvotes · 0 comments - [AI red teaming comes of age](https://daily.dev/posts/ai-red-teaming-comes-of-age-flzm3plsq) · CSO Online · 0 upvotes · 0 comments - [Why AI Red Teaming Is the Next Step After AI-SPM](https://daily.dev/posts/why-ai-red-teaming-is-the-next-step-after-ai-spm-lui1h2coa) · Snyk · 1 upvotes · 0 comments - [Cisco AI Defense: Explorer Edition Brings Agentic AI Red Teaming to Builders](https://daily.dev/posts/cisco-ai-defense-explorer-edition-brings-agentic-ai-red-teaming-to-builders-csta5nd0m) · Cisco · 0 upvotes · 0 comments --- Tags: [#security](https://daily.dev/tags/security), [#mcp](https://daily.dev/tags/mcp), [#prompt-injection](https://daily.dev/tags/prompt-injection) [View this post on daily.dev](https://daily.dev/posts/launching-ai-red-teaming-pentest-your-ai-agents-and-mcp-servers-all-alongside-your-app-d2b9jsftc) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Launching AI Red Teaming: pentest your AI agents and MCP servers all alongside your app","url":"https://daily.dev/posts/launching-ai-red-teaming-pentest-your-ai-agents-and-mcp-servers-all-alongside-your-app-d2b9jsftc","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/launching-ai-red-teaming-pentest-your-ai-agents-and-mcp-servers-all-alongside-your-app-d2b9jsftc"},"datePublished":"2026-10-21T09:13:41.419Z","dateModified":"2026-10-21T09:14:46.730Z","description":"Escape has launched AI Red Teaming, a new capability inside its AI Pentesting product, live for customers starting October 21. It tests AI agents and MCP...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/08cc1c7ee7e96a63e9daca3db13c8153?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/08cc1c7ee7e96a63e9daca3db13c8153?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Security Boulevard","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Security Boulevard","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/3613c832180040de8d85bb29f74395be","url":"https://daily.dev/sources/securityboulevard"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/launching-ai-red-teaming-pentest-your-ai-agents-and-mcp-servers-all-alongside-your-app-d2b9jsftc","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"security,mcp,prompt-injection","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Security Boulevard","item":"https://daily.dev/sources/securityboulevard"},{"@type":"ListItem","position":3,"name":"Launching AI Red Teaming: pentest your AI agents and MCP servers all alongside your app"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/launching-ai-red-teaming-pentest-your-ai-agents-and-mcp-servers-all-alongside-your-app-d2b9jsftc#faq","mainEntity":[{"@type":"Question","name":"How can I pentest an MCP server for vulnerabilities?","acceptedAnswer":{"@type":"Answer","text":"Escape's AI Red Teaming, part of its AI Pentesting product, lists an MCP server's tools and attacks them over MCP's own JSON-RPC protocol, which REST or GraphQL scanners can't enumerate or fuzz. It can also reach the MCP server indirectly by attacking it through the chatbot that fronts it, testing the real attack path rather than the server in isolation. Teams rolling out MCP servers can follow security tooling coverage like this on daily.dev."}},{"@type":"Question","name":"What is multi-turn jailbreaking in AI security testing?","acceptedAnswer":{"@type":"Answer","text":"Multi-turn jailbreaking refers to vulnerabilities in a chatbot's guardrails that only appear after several exchanges build up conversational context, rather than from a single malicious prompt. Escape's Cascade tool holds extended conversations with a chatbot, adapting to each response and pushing further to surface these jailbreaks, then records the full conversation as reproducible evidence. Developers hardening chatbot guardrails can track techniques like this via daily.dev."}},{"@type":"Question","name":"How do you prove a prompt injection attack is actually exploitable?","acceptedAnswer":{"@type":"Answer","text":"Exploitability is proven by checking whether the injection achieves something concrete, such as an unauthorized tool call, another tenant's data, or a privileged action, rather than just an unexpected chatbot reply. Escape's AI Red Teaming attaches proof of the exploit to findings, which is meant to end debate between security and engineering over how severe a given prompt injection really is. Anyone triaging AI security findings can follow proof-based approaches like this on daily.dev."}}]} ```