https://cybertoolchain.io/tools/zenity New releases and features in Zenity, tracked by The Cyber Toolchain. en Sat, 17 Oct 2026 07:52:37 GMT https://zenity.io/blog?category=product-announcements https://zenity.io/blog?category=product-announcements Sat, 17 Oct 2026 07:52:37 GMT Zenity adds Universal Runtime Control for Claude Enterprise via Anthropic's Inference Hooks to validate tool calls before execution.⁠︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄⁠ • Supports Anthropic's Inference Hooks to validate Claude Enterprise tool calls at runtime before they execute.⁠︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄⁠ Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/zenity-launches-universal-runtime-control-for-claude-enterprise https://zenity.io/blog/zenity-launches-universal-runtime-control-for-claude-enterprise Fri, 16 Oct 2026 07:47:43 GMT Zenity adds pre-execution tool-call enforcement for Claude Enterprise via Anthropic's Inference Hooks, covering all Claude surfaces with no endpoint agent required.⁠︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄⁠ • Enables pre-execution allow-or-deny verdicts on Claude tool calls via Anthropic's `tool_call` Inference Hooks event, blocking prohibited operations before they run across claude.ai (web, desktop, mobile), Cowork, Claude Code, Claude for Microsoft 365, Claude in Chrome, Claude's built-in browser, and Claude Tag in Slack — with no endpoint agent or browser extension required.⁠︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄⁠ • Introduces Zenity Boundaries read/write classification for supported tool calls, letting policy allow reads while blocking deletes, bulk updates, external shares, shell commands, or form submissions that violate policy — without depending on tool names that vary across connectors. • Adds tool provenance enforcement: Boundaries can restrict tool calls to verified MCP servers and connectors, or block calls from unverified or local origins, using provider metadata supplied with each `tool_call` event. • Adds prompt-injection-tainted session blocking: when Zenity detects an injected instruction in content Claude reads, it marks the session as tainted and blocks any subsequent tool call that acts on that instruction, with the denial reason surfaced in Claude and the event chain available in Zenity's Compliance API audit log. • Surfaces tool-call denials in Claude's Activity Feed with a reference to the associated finding, giving security teams a reviewable chain of events per blocked action. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/zenity-launches-universal-runtime-control-for-claude-enterprise https://zenity.io/blog/zenity-launches-universal-runtime-control-for-claude-enterprise Fri, 16 Oct 2026 07:47:43 GMT Zenity adds pre-execution tool-call enforcement for Claude Enterprise via Anthropic's Inference Hooks, covering all Claude surfaces with no endpoint agent required.⁠︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄⁠ • Enables pre-execution allow-or-deny verdicts on Claude tool calls via Anthropic's `tool_call` Inference Hooks event, blocking prohibited operations before they run across claude.ai (web, desktop, mobile), Cowork, Claude Code, Claude for Microsoft 365, Claude in Chrome, Claude's built-in browser, and Claude Tag in Slack — with no endpoint agent or browser extension required.⁠︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄⁠ • Introduces Zenity Boundaries read/write classification for supported tool calls, letting policy allow reads while blocking deletes, bulk updates, external shares, shell commands, or form submissions that violate policy — without depending on tool names that vary across connectors. • Adds tool provenance enforcement: Boundaries can restrict tool calls to verified MCP servers and connectors, or block calls from unverified or local origins, using provider metadata supplied with each `tool_call` event. • Adds prompt-injection-tainted session blocking: when Zenity detects an injected instruction in content Claude reads, it marks the session as tainted and blocks any subsequent tool call that acts on that instruction, with the denial reason surfaced in Claude and the event chain available in Zenity's Compliance API audit log. • Surfaces tool-call denials in Claude's Activity Feed with a reference to the associated finding, giving security teams a reviewable chain of events per blocked action. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/august-2026-product-updates https://zenity.io/blog/august-2026-product-updates Thu, 17 Sep 2026 07:33:56 GMT Zenity adds Blue Agent triage, a Boundaries CLI, OTel/gen_ai span ingestion, and GA coverage for GitHub Copilot, OpenAI Codex, and Microsoft Foundry.⁠︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄⁠ • Adds a Boundaries `CLI` that lets teams manage boundaries, custom taints, and reusable values as version-controlled files, with coding agents like Claude Code and Codex authoring policy against Zenity's schema.⁠︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄⁠ • Adds a `Custom Agents` integration that ingests OpenTelemetry (`OTel`) and `gen_ai` spans for visibility and inline enforcement via the `Evaluate API`, connecting any agentic application to Zenity without a dedicated connector. • Introduces `Blue Agent`, the first Guardian Agent, which triages every new finding in suggest mode — returning a verdict (benign true positive, false positive, or true positive) with reasoning, confidence level, and full trace attached. • Adds system taints that classify tool calls as `read` or mutative operations (`write`, `update`, `delete`, `execute`), enabling rules based on what an action does rather than command strings or tool names. • Adds `Prevent` and `Detect` labels to every taint in the Boundaries manager so teams can see at a glance whether a rule blocks in real time or flags after the fact. • Extends the Boundaries AI assistant to create required custom taints while drafting a rule and to run draft rules against recent production activity on command before applying. • Brings GitHub Copilot and OpenAI Codex to general availability with inventory, activity monitoring, and inline prevention served through hooks. • Adds full coverage for Microsoft Foundry (the rebuilt agent stack that replaced Azure AI Foundry at Ignite 2025), including inventory, graph, runtime activity, and AI detection and response (`AIDR`) findings. • Surfaces Zenity runtime risk signals directly inside Microsoft Agent 365, alongside Entra, Purview, and Defender. • Enables near-real-time audit log event streaming to Splunk or Microsoft Sentinel. • Publishes step-by-step integration guides for Cribl, LiteLLM, and Kong to connect agent activity to the Zenity portal. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/openapi.json https://zenity.io/openapi.json Wed, 16 Sep 2026 07:47:26 GMT Zenity now publishes an API — 3 endpoints across 3 areas: .well Known, Api, Markdown⁠︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄⁠ • .well Known (1 endpoint) — read⁠︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄⁠ • Api (1 endpoint) — read • Markdown (1 endpoint) — read Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/claude-compliance-api-integration https://zenity.io/blog/claude-compliance-api-integration Fri, 12 Jun 2026 00:00:00 GMT Zenity integrates with Claude Compliance API for post-session correlation and inline prevention of prompt injection and exfiltration. • Integrates with `Claude's Compliance API` to bring Claude activity into Zenity's AI security and governance platform • Provides session replay of prompts, responses, tool interactions, and execution context for Claude agents • Correlates agent sessions in Claude Code with pull requests, commits, and code changes • Maintains an inventory of MCP servers, skills, plugins, and connected extensions across the Claude estate • Evaluates configuration posture across managed, user, project, and local scopes before sessions begin • Enforces organization-level policy for MCP servers, skills, plugins, and other agent extensions (allow, block, review) • Detects embedded credentials, suspicious payloads, and injection patterns in configuration and extension layers • Provides real-time AIDR detection of prompt injection across user, AI, and tool planes • Blocks destructive actions that exceed the intended task before they reach source code or secrets • Detects memory poisoning, behavioral drift, and C2/exfiltration attempts through agent prompts and tool calls • Preserves structured evidence of Claude activity for compliance teams to map to audit requirements Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/zenity-aws-security-hub https://zenity.io/blog/zenity-aws-security-hub Wed, 20 May 2026 00:00:00 GMT Zenity integrates with AWS Security Hub Extended to unify AI agent findings into the OCSF-based risk correlation engine. • AI agent posture violations, runtime anomalies, and detected threats appear in AWS Security Hub Extended via Open Cybersecurity Schema Framework (OCSF) Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/securing-the-ai-that-runs-the-enterprise-zenity-servicenow-secops https://zenity.io/blog/securing-the-ai-that-runs-the-enterprise-zenity-servicenow-secops Tue, 24 Mar 2026 00:00:00 GMT Zenity adds native AI Security Posture Management and agent inventory integration with ServiceNow SecOps. • Adds native AI Security Posture Management (AISPM) engine to Zenity, now integrated into ServiceNow SecOps for continuous analysis of agents' excessive permissions, unsafe prompt logic, data exposure risks, and broken integrations. • Extends Zenity's agent inventory visibility to include underlying workflows, API dependencies, identities, credentials, and data flows for agents across non-ServiceNow environments like SaaS and endpoints. • Enables real-time vulnerability insights, misconfiguration detection, and risk scoring for AI agents within ServiceNow SecOps workflows. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/open-source-security-openclaw https://zenity.io/blog/open-source-security-openclaw Mon, 23 Mar 2026 00:00:00 GMT Zenity releases open-source security infrastructure for OpenClaw agent workflows • Introduces a new open-source security framework for Zenity's OpenClaw that allows teams to add detection and blocking capabilities directly into agent execution flows. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/seeing-what-ai-touches-introducing-data-lens https://zenity.io/blog/seeing-what-ai-touches-introducing-data-lens Wed, 21 Jan 2026 00:00:00 GMT Introducing Data Lens for AI agent runtime data visibility and oversight. • Launches 'Data Lens', a new capability within Zenity AI Observability that aggregates AI activity to show which data items agents touch, how frequently, and to whom at runtime. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/safe-harbor-an-open-source-abort-mission-button-for-your-ai-agent https://zenity.io/blog/safe-harbor-an-open-source-abort-mission-button-for-your-ai-agent Thu, 04 Dec 2025 00:00:00 GMT Zenity releases Safe Harbor, an open-source wrapper for OpenAI, Anthropic, and Gemini clients that adds a built-in 'safe action' for agents to abort unsafe workflows. • Adds `Safe Harbor` wrapper for OpenAI, Anthropic, or Google Gemini clients to provide agents with a dedicated safe action for aborting malicious instructions Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/your-browser-is-becoming-an-agent-zenity-keeps-it-from-becoming-a-threat https://zenity.io/blog/your-browser-is-becoming-an-agent-zenity-keeps-it-from-becoming-a-threat Thu, 04 Dec 2025 00:00:00 GMT Zenity adds detection and governance for agentic browsers like ATLAS, Comet, and Dia. • Adds automatic discovery of agentic browsers (ATLAS, Comet, Dia, MCPs) on managed and unmanaged devices. • Enables real-time guardrails to flag or block unauthorized actions by agentic browsers. • Monitors agentic browser behavior for sensitive data leakage including secrets, PII, PCI, and PHI. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/zenity-brings-inline-prevention-to-microsoft-foundry-and-copilot-studio https://zenity.io/blog/zenity-brings-inline-prevention-to-microsoft-foundry-and-copilot-studio Tue, 18 Nov 2025 00:00:00 GMT Zenity brings inline prevention to Microsoft Foundry (preview) and Copilot Studio agents. • Adds `inline prevention` capability for agents built in Microsoft Foundry (preview) • Adds `inline prevention` capability for agents built in Microsoft Copilot Studio Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/closing-the-guardrail-gap-runtime-protection-for-openai-agentkit https://zenity.io/blog/closing-the-guardrail-gap-runtime-protection-for-openai-agentkit Tue, 04 Nov 2025 00:00:00 GMT Zenity adds runtime protection for OpenAI AgentKit with deterministic enforcement of data leakage, secrets, and unsafe actions. • Introduces runtime protection capability that inspects every interaction between users and AgentKit-based agents to enforce policy before risk turns into exposure • Adds intent-focused inspection engine that analyzes underlying action intent rather than simple prompt-and-response to detect manipulative behavior hiding behind safe language • Deploys rule-based, deterministic enforcement for three specific categories: Data Leakage Detection (blocking exfiltration in responses or tool outputs), Secrets Exposure Prevention (stopping embedded API keys/credentials/tokens), and Unsafe Response Blocking (preventing policy-violating actions) Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/security-for-autonomous-agents-and-reducing-shadow-ai https://zenity.io/blog/security-for-autonomous-agents-and-reducing-shadow-ai Fri, 19 Sep 2025 00:00:00 GMT Zenity expands its AISPM platform with Agent Triggers and governance tools for autonomous AI workflows. • Zenity expands its AISPM platform with Agent Triggers and governance tools for autonomous AI workflows. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/zenity-and-microsoft-copilot-studio-deliver-inline-protection https://zenity.io/blog/zenity-and-microsoft-copilot-studio-deliver-inline-protection Mon, 08 Sep 2025 00:00:00 GMT Zenity adds inline protection integration with Microsoft Copilot Studio • Adds `Microsoft Copilot Studio` integration to intercept agent tool invocations in real time • Enables detection of indirect prompt injection attacks via intent analysis on agent behavior Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/securing-the-ai-agent-era-across-saas-endpoint-and-cloud https://zenity.io/blog/securing-the-ai-agent-era-across-saas-endpoint-and-cloud Tue, 05 Aug 2025 00:00:00 GMT Zenity unifies governance and runtime protection for SaaS, device-based, and home-grown AI agents • Enables real-time policy enforcement for SaaS-managed agents (Copilot Studio, Salesforce Agentforce) to prevent data exfiltration and misuse of internal APIs. • Extends agent-centric discovery to endpoints to identify device-based agents (GitHub Copilot, Cursor), connected Model Context Protocol servers, and their interactions. • Provides cross-cloud inventory and governance for home-grown agents across AWS Bedrock, Azure AI Foundry, and Google Cloud to flag misconfigurations and risky deployments. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/zenity-and-microsoft-copilot-studio-extend-ai-agent-security-from-buildtime-to-runtime https://zenity.io/blog/zenity-and-microsoft-copilot-studio-extend-ai-agent-security-from-buildtime-to-runtime Thu, 22 May 2025 00:00:00 GMT Zenity extends Microsoft Copilot Studio integration to provide runtime AISPM and AIDR for AI Agents • Adds `AISPM` (AI Security Posture Management) capabilities for agents built in Microsoft Copilot Studio • Adds `AIDR` (AI Detection & Response) for real-time threat detection including prompt injection and data leakage during agent runtime • Extends Zenity integration with Microsoft Copilot Studio to provide continuous visibility into agent activity, usage, and access Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/introducing-zenity-for-chatgpt-enterprise https://zenity.io/blog/introducing-zenity-for-chatgpt-enterprise Wed, 23 Apr 2025 00:00:00 GMT Zenity launches for ChatGPT Enterprise, adding agent-level visibility and governance. • Adds integration with OpenAI Compliance API to bring Zenity to ChatGPT Enterprise • Enables buildtime scanning of Custom GPTs for misconfigurations, hardcoded secrets, and excessive permissions • Provides runtime threat detection for prompt injection, RAG manipulation, and unauthorized data access Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity https://zenity.io/blog/securing-enterprise-copilots-a-fresh-and-agent-less-application-security-approach https://zenity.io/blog/securing-enterprise-copilots-a-fresh-and-agent-less-application-security-approach Fri, 20 Sep 2024 00:00:00 GMT Zenity launches AI Trust Layer providing agentless security and governance for Microsoft 365 Copilot. • Adds visibility into how enterprise copilots and agents are used to identify real-time risks such as sensitive data exposure. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/zenity