SAN FRANCISCO, March 23, 2026 (GLOBE NEWSWIRE) — Snyk, the AI safety corporate, these days introduced the release of its new Agent Safety answer and the Basic Availability (GA) of Snyk Evo AI-SPM. The answer arrives as enterprises hastily deploy self reliant coding brokers – like Claude Code, Cursor, and Devin – ceaselessly with out a safety structure governing what those methods construct, the gear they invoke, or the movements they take throughout manufacturing infrastructure.
Popping out of stealth and attaining GA, Snyk Evo AI-SPM – incubated throughout the corporate’s devoted AI innovation arm – expands the operational functions of the AI Safety Material. This enforcement structure governs self reliant brokers from the instant they’re presented into device via each and every motion they absorb manufacturing. As enterprises deploy self reliant brokers throughout their device stack, Snyk supplies the enforcement layer that each and every agentic workflow should move via to be enterprise-safe.
The Shadow AI Disaster and the Governance GapAutonomous device brokers are actually writing, enhancing, and deploying code at mechanical device pace, and conventional safety overview processes can’t stay tempo. Snyk’s 2026 State of Agentic AI Adoption Document presentations that enterprises that believed that they had AI underneath keep watch over have came upon they don’t. For each and every AI fashion deployed, enterprises introduce just about thrice as many untracked device elements.
Throughout early get entry to, throughout greater than 500 Evo scans, organizations with current cloud safety and CNAPP controls in position had been stunned to find ungoverned agentic AI elements of their codebases that had in the past bypassed their safety stack.
Cloud safety platforms display the place AI runs after deployment. Snyk presentations the place AI is presented into device and enforces governance insurance policies sooner than it ever reaches the cloud.
“Agentic architectures turn governance into a software supply chain problem,” mentioned Manoj Nair, Leader Innovation Officer at Snyk. “Our value is confirming which findings are real and exploitable, using ground truth data from a decade of enterprise deployment that no AI model can produce alone. Claude finds. Snyk confirms. The agent fixes only what’s real.”
For early adopters like WEX, a world chief in intelligence-led cost and workflow answers, gaining this speedy working out in their AI footprint has been a essential first step.
“One of the things that was exciting about Evo AI-SPM is the visibility. It only took an afternoon to set it up and less time to pull a report and have full visibility,” mentioned Jason Langston, Director, Product Safety at WEX. “Here at WEX, we take a business-impact approach to AI. Being able to put our arms around the full breadth of what was actually in place was a super helpful foundation to start from.”
Evo AI-SPM: Turning Governance into Enforceable ControlsEvo AI-SPM serves because the engine for Snyk’s Agent Safety answer, operationalizing AI safety via specialised, automatic brokers. To do away with guide governance, a Discovery Agent robotically maps the “code-first” assault floor to generate a reside AI-BOM. This stock is frequently enriched by means of a Possibility Intelligence Agent with metadata, hallucination and bias metrics, and contextual safety indicators. In spite of everything, a Coverage Agent interprets plain-English governance intent into machine-enforceable safety guardrails that execute natively all through CI pipelines.
Governing the Complete Agent Lifecycle: Atmosphere, Artifact, and BehaviorTo permit builders to leverage robust AI coding gear with out compromising their safety posture, Snyk Agent Safety delivers a unified structure that secures the 3 essential levels of agentic construction:
Atmosphere — Agent Scan (Open Preview): Secures the provision chain of gear brokers depend on by means of making sure each and every MCP server and agent ability is understood, depended on, and ruled.Artifact — Snyk Studio: Enforces safety validation at once throughout the CI/CD pipeline as code is produced. Studio is already deployed throughout 300+ venture consumers and natively built-in into Claude Code, Cursor, and Devin workflows.Habits — Agent Guard (Non-public Preview): Supplies real-time enforcement throughout the construction loop, preventing damaging instructions and governing precisely how brokers function.
Past construction, Snyk could also be securing AI-native programs at runtime. Within the realm of dynamic trying out, Snyk API & Internet objectives authorization and industry good judgment vulnerabilities—comparable to BOLA and IDOR which can be often presented by means of AI-generated code and pose important chance when uncovered to self reliant brokers.
Moreover, Agent Crimson Teaming (Open Preview) deploys self reliant brokers to simulate multi-turn assault flows, frequently exposing vulnerabilities so they are able to be fastened sooner than they’re exploited in manufacturing.
Availability
Evo AI-SPM — together with the Discovery Agent, Possibility Intelligence Agent, and Coverage Agent — and API & Internet Trying out are each Most often To be had (GA) these days.Snyk Studio is now embedded at once into AI-driven construction workflows throughout the Snyk platform.Agent Scan and Agent Crimson Teaming are these days to be had in Open Preview, whilst Agent Guard is to be had in Non-public Preview. Snyk is actively in the hunt for design companions for its preview methods.
Snyk is showcasing its new Agent Safety answer and the GA functions of Evo AI-SPM this week on the RSA Convention 2026. To be informed extra or request a demo, seek advice from https://snyk.io/occasions/rsac/.
About Snyk Snyk, the AI safety corporate, empowers the AI-driven venture to expand and protected its long run, making sure organizations can consider AI to innovate with out limits. The Snyk AI Safety Platform serves because the business’s AI Safety Material, weaving coverage at once into the drift of advent to protected GenAI code, AI-native programs, and agentic methods. By means of handing over visibility, keep watch over, and self reliant protection protected at inception, Snyk allows over 4,800 world consumers to construct fearlessly within the AI generation.
