llms.txt
Compare with other sites' llms.txt files →# Agen by Frontegg
> Agen.co is the identity-native agentic governance platform from Frontegg. It lets enterprises become AI-native securely: every AI agent is anchored to the identity and the human behind it, and every action is governed at runtime, on one platform across endpoint, browser, gateway and cloud.
Agen covers internal agents (Copilot, coding agents, MCP clients used by employees) and customer-facing agents built into SaaS products. Key areas: discovery of shadow AI and agents, per-action policy enforcement with step-up and human-in-the-loop approval, an MCP gateway with a catalog of governed integrations, and identity for agents built on Frontegg's CIAM platform.
## Product
- [MCP Gateway](https://agen.co/mcp-gateway)
- [AI Agent Observability to Stop Shadow AI](https://agen.co/platform/observability): Catalog every AI agent, monitor user-to-agent activity, observe MCP usage, and track skills. Tracking agentic access is the only way to stop Shadow AI.
- [Identity-Native AI Agent Governance Platform](https://agen.co/platform): The identity-native agentic governance platform: discover every agent, broker every connection, govern every action at runtime, and prove it happened.
- [AI Agent Discovery: Find Shadow AI in Your Org](https://agen.co/discover): Find the AI agents your people already connected, including shadow AI nobody approved. See who authorized each one, what it can access, and its risk.
- [Govern Every AI Agent Action at Runtime](https://agen.co/govern): Govern what every AI agent does the moment it acts, with per-action decisions, step-up, and human-in-the-loop anchored to the identity behind it.
- [Policy engine](https://agen.co/policy-engine)
- [Shield: Runtime Protection for AI Agents](https://agen.co/shield): Shield stops risky AI agent actions as they happen across endpoint, browser, gateway, and cloud, with one policy enforced identically everywhere.
- [Identity Foundation for Every AI Agent](https://agen.co/identity): Anchor every AI agent to the identity behind it, with machine identities, secretless access, your existing IdP, and a seven-year identity foundation.
- [Install Agen MCP Gateway in Claude, ChatGPT, Cursor](https://agen.co/mcp-gateway/install): Install the Agen.co MCP Gateway in minutes. Connect your AI agents to enterprise tools with built-in governance, audit, and access control.
- [Govern Claude Code at Runtime, Per Action](https://agen.co/govern/claude-code): How Agen governs Claude Code: every file, shell and MCP action judged at runtime against the developer behind the agent. Engineering keeps the tool.
- [Govern OpenAI Codex at Runtime, Per Action](https://agen.co/govern/codex): How Agen governs Codex on developer laptops and in cloud tasks: every action judged at runtime against the developer behind it, with evidence at action time.
- [Govern Cursor Agent Mode at Runtime, Per Action](https://agen.co/govern/cursor): How Agen governs Cursor agent mode, auto-run commands, MCP servers and background agents: judged per action at runtime, with evidence at action time.
- [Govern Windsurf Cascade at Runtime, Per Action](https://agen.co/govern/windsurf): How Agen governs Windsurf Cascade: file edits, terminal commands and MCP calls judged per action at runtime against the developer behind them.
- [Govern GitHub Copilot Agent Mode and the Coding Agent](https://agen.co/govern/github-copilot): How Agen governs GitHub Copilot agent mode and the coding agent in Actions: every action judged at runtime against the developer behind it.
- [Govern ChatGPT: Stop the Paste, Govern the Action](https://agen.co/govern/chatgpt): How Agen governs ChatGPT: BrowserShield stops company data at the paste, every account and custom GPT becomes visible, GPT actions are decided per action.
- [Govern Claude: Stop the Paste, Govern the Connector](https://agen.co/govern/claude): How Agen governs Claude: BrowserShield stops company data at the paste, every account and connector becomes visible, MCP connectors are decided per call.
- [Govern Gemini: Stop the Paste, Govern the Agents](https://agen.co/govern/gemini): How Agen governs Gemini: BrowserShield stops company data at the paste, Gemini Enterprise agents are decided at the gateway, Gemini CLI on the device.
- [Govern Microsoft 365 Copilot Agents Per Action](https://agen.co/govern/microsoft-365-copilot): How Agen governs Microsoft 365 Copilot: BrowserShield stops company data at the paste, and agents built on Copilot are decided at the gateway per action.
- [Govern Copilot Studio Agents Per Action at Runtime](https://agen.co/govern/copilot-studio): How Agen governs Copilot Studio agents acting through Power Platform connectors and flows: judged per action at runtime with a named owner behind every one.
- [Govern Agentforce Agents Per Action at Runtime](https://agen.co/govern/agentforce): How Agen governs Agentforce agents acting on CRM records, flows, Apex and MuleSoft APIs: judged per action at runtime with a named owner behind every one.
## Solutions and use cases
- [AI Agent Accountability and Audit Trails](https://agen.co/solutions/accountability): Tie every AI agent to a named human owner, with per-action audit trails anchored to identity and generated at action time for real accountability.
- [Confident AI Adoption With Security in Loop](https://agen.co/solutions/adoption): Adopt AI at full speed with security in the loop on every action. Get approvals in hours, guardrails per action, and shadow AI surfaced across the org.
- [Autonomous AI Governance Operations](https://agen.co/solutions/autonomous-operations): Built-in Agen.co agents run tedious governance work autonomously: policy drafting, compliance mapping, and step-up triage, so the manual queue trends to zero.
- [AI Agent Risk Prevention at Runtime](https://agen.co/solutions/risk): Block the malicious AI agent action before it executes, with runtime verdicts, least privilege by default, blast-radius scoring, and device enforcement.
- [AI Agent Governance Use Cases](https://agen.co/use-cases): See how enterprises use Agen.co to govern AI agents, internal and customer-facing, on one identity fabric with a named human accountable for each.
- [Agen for Work](https://agen.co/use-cases/agen-for-work)
- [Agen for SaaS](https://agen.co/use-cases/agen-for-saas)
- [AI Agent Governance by Industry](https://agen.co/industries): Agentic governance for financial services, software, healthcare, and consumer companies. Different stakes across sectors, held to one governance thread.
- [AI Agent Governance for Every Security Role](https://agen.co/personas): Agentic governance for the CISO, security team, IT and platform, and developers. Four roles working from one governed answer across every AI agent.
- [Copilot governance](https://agen.co/lp/copilot-governance)
- [Coding agent governance](https://agen.co/lp/coding-agent-governance)
## Comparisons
- [Agent Governance Comparisons](https://agen.co/compare): Capability-by-capability comparisons of agent-governance platforms, scored 0–5 against vendor documentation with every source published — including the rows we lose.
- [Aembit vs Agen.co: Compared](https://agen.co/compare/aembit): Aembit vs Agen.co across 16 scored capabilities: runtime enforcement, identity, coverage and cost, with the four rows that tie and the two Aembit wins.
- [Airia vs Agen.co: Compared](https://agen.co/compare/airia): How Airia builds, runs and governs AI agents — Discover, Secure, Govern — scored against Agen.co across 16 capabilities. Runtime enforcement, accountability, coverage and cost, with sources.
- [Astrix Security vs Agen.co: Compared](https://agen.co/compare/astrix-security): Astrix Security vs Agen.co: agent discovery, ownership, runtime enforcement and cost across 16 scored capabilities, with four ties and the one Astrix wins.
- [Entro Security vs Agen.co: Compared](https://agen.co/compare/entro-security): Entro Security vs Agen.co: secrets, non-human identities and agent governance scored across 16 capabilities, with sources and the three rows Entro wins.
- [Eve Security vs Agen.co: Compared](https://agen.co/compare/eve-security): Eve Security vs Agen.co: Discover, Govern, Enforce and Agent-in-the-Loop scored across 16 capabilities, with sources and the rows Eve Security wins.
- [Highflame vs Agen.co: Compared](https://agen.co/compare/highflame): Highflame vs Agen.co: ZeroID, Cedar policy and per-action authorization scored across 16 capabilities, with sources and the four rows Highflame wins.
- [Keycard vs Agen.co: Compared](https://agen.co/compare/keycard): How Keycard issues task-scoped agent credentials from Cedar policy, scored against Agen.co on 16 capabilities, with five tied rows and the one Keycard wins.
- [Lasso Security vs Agen.co: Compared](https://agen.co/compare/lasso): How Lasso Security enforces AI policy inline at the gateway, scored against Agen.co on 16 capabilities, with sources, four level rows and the three Lasso wins.
- [Microsoft Agent Governance vs Agen.co: Compared](https://agen.co/compare/microsoft-agent-governance): How Microsoft governs agents — Agent 365, Entra Agent ID, Purview, Defender — scored against Agen.co across 28 capabilities. Runtime enforcement, accountability, coverage, and cost, with sources.
- [Noma Security vs Agen.co: Compared](https://agen.co/compare/noma-security): How Noma Security discovers and protects AI agents — AISPM, red teaming, runtime protection — scored against Agen.co across 16 capabilities. Enforcement, accountability, coverage and cost, with sources.
- [Oasis Security vs Agen.co: Compared](https://agen.co/compare/oasis-security): How Oasis Security governs non-human identities and intent-aware agent access, scored against Agen.co on 16 capabilities, with the three rows Oasis wins.
- [Obsidian Security vs Agen.co: Compared](https://agen.co/compare/obsidian-security): How Obsidian Security discovers AI agents and blocks via platform callbacks, scored against Agen.co on 16 capabilities, with sources and the two rows it wins.
- [Agen.co vs Onyx Security — Agent Governance Compared](https://agen.co/compare/onyx-security): Onyx Security enforces inline across five AI surfaces with five verdict types. See 16 capabilities scored 0–5 against vendor documentation — with every source, the seven rows that come out level, and the two rows Onyx wins.
- [Prompt Security vs Agen.co: Compared](https://agen.co/compare/prompt-security): Prompt Security's browser, endpoint and MCP sensors scored against Agen.co on 16 capabilities, 0–5, with sources and the three rows Prompt Security wins.
- [Straiker vs Agen.co: Compared](https://agen.co/compare/straiker): How Straiker secures AI agents — Discover, Ascend, Defend — scored against Agen.co across 16 capabilities. Runtime enforcement, accountability, coverage and cost, with sources.
- [Agen.co vs Ultra Security — Agent Governance Compared](https://agen.co/compare/ultra-security): Ultra Security governs MCP tool calls inline, on the host. See 16 capabilities scored 0–5 against vendor documentation — with every source, the six rows that come out level, and the row Ultra wins.
- [WitnessAI vs Agen.co: Compared](https://agen.co/compare/witnessai): How WitnessAI governs AI and agents — Observe, Control, Protect — scored against Agen.co across 16 capabilities. Runtime enforcement, accountability, coverage, and cost, with sources.
- [Zenity vs Agen.co: Compared](https://agen.co/compare/zenity): How Zenity secures AI agents — Surface, Enforce, Protect — scored against Agen.co across 16 capabilities. Runtime enforcement, accountability, coverage, and cost, with sources.
- [Aegis Security vs Agen.co: Compared](https://agen.co/compare/aegis-security): How Aegis Security discovers and enforces on AI agents, scored against Agen.co on 14 capabilities: runtime, accountability, coverage and operations.
- [Gray Swan vs Agen.co: Compared](https://agen.co/compare/gray-swan): Gray Swan — Cygnal, Shade, Arena — scored against Agen.co on 16 capabilities: runtime enforcement, accountability, coverage and deployment, with sources.
- [Lema vs Agen.co: Compared](https://agen.co/compare/lema): Lema governs the AI inside your vendors. Agen.co governs what your agents do. Fourteen capabilities scored side by side, including the rows Lema wins.
- [Lumia vs Agen.co: Compared](https://agen.co/compare/lumia): Lumia Security governs AI use from the network layer. Scored against Agen.co on 16 capabilities: runtime, identity, coverage and operations, with sources.
- [QuilrAI vs Agen.co: Compared](https://agen.co/compare/quilr): How QuilrAI governs AI agents through its LLM gateway, MCP gateway and endpoint agent, scored against Agen.co across 16 capabilities, with sources.
- [Runlayer vs Agen.co: Compared](https://agen.co/compare/runlayer): How Runlayer governs AI tool calls at its MCP gateway, scored against Agen.co on 15 capabilities: runtime, accountability, coverage and cost, with sources.
- [Solo.io vs Agen.co: Compared](https://agen.co/compare/solo-io): How Solo.io's agentgateway, ambient mesh, kagent and agentregistry govern AI agents, scored against Agen.co on 16 capabilities, with sources.
- [Virtue AI vs Agen.co: Compared](https://agen.co/compare/virtue-ai): How Virtue AI secures AI agents — ActionGuard, MCPGuard, red teaming — scored against Agen.co across 15 capabilities, with sources and the rows it wins.
- [Wiz vs Agen.co: Compared](https://agen.co/compare/wiz): Wiz for AI security, scored: AI-SPM, AI-BOM and Wiz Defend against Agen.co on 16 capabilities, with sources and the four rows Wiz wins.
## Learning Center
- [Learning Center](https://agen.co/learning-center)
- [Agentic Risk Map: How to Map & Score Agent Risk](https://agen.co/learning-center/agentic-risk-map): An agentic risk map is a reusable framework for inventorying, scoring, and containing AI agent risk. Learn the risk dimensions, scoring rubric, and build steps.
- [What Is LLM Observability? Pillars, Metrics, Architecture](https://agen.co/learning-center/llm-observability): LLM observability lets teams trace, monitor, and evaluate LLM apps in production. Learn the three pillars, the metrics that matter, and best practices.
- [AI Security Posture Management (AISPM): A Complete Guide](https://agen.co/learning-center/ai-security-posture-management): AI security posture management (AISPM) finds and fixes risk across models, agents, and pipelines. See how it works, how it differs from CSPM, and how to start.
- [Shadow AI: Risks, Examples & How to Govern It](https://agen.co/learning-center/what-is-shadow-ai): Shadow AI is the unsanctioned use of AI tools, agents, and MCP servers inside your org. Learn the real risks, examples, and how to detect and govern it.
- [AI Red Teaming: A Complete Guide to Testing AI & LLMs](https://agen.co/learning-center/ai-red-teaming): What is AI red teaming? How adversarial testing exposes LLM and agent flaws before attackers do, with frameworks from OWASP, MITRE ATLAS, and NIST AI RMF.
- [Enterprise MCP Gateway Requirements: A Selection Framework](https://agen.co/learning-center/enterprise-mcp-gateway): What makes an MCP gateway enterprise-grade? Evaluate identity, authorization, and governance with a vendor-neutral checklist and a build-vs-buy verdict.
- [EU AI Act Compliance: Requirements & 2026 Deadlines](https://agen.co/learning-center/eu-ai-act-compliance): EU AI Act compliance explained: risk tiers, provider and deployer obligations, GPAI rules, 2026 deadlines, fines up to 7% of turnover, and where to start.
- [AI Agent Workforce Management: Complete Guide](https://agen.co/learning-center/ai-agent-workforce-management): AI agent workforce management is how enterprises onboard, govern, secure, and oversee a fleet of autonomous AI agents. Learn the lifecycle and control plane.
- [AI Guardrails: Types, Architecture & Best Practices](https://agen.co/learning-center/ai-guardrails): AI guardrails are runtime controls on what an AI agent can take in, output, and do. See the types, architecture, and best practices that keep agents in check.
- [MCP Server Hosting: Self-Host vs Managed Guide](https://agen.co/learning-center/mcp-server-hosting): MCP server hosting is an identity and policy problem, not just compute. Compare self-hosted vs managed vs gateway-fronted models and what production demands.
- [AI Threat Detection for AI Agents: The Complete Guide](https://agen.co/learning-center/ai-threat-detection): AI threat detection spots and contains rogue or compromised AI-agent behavior at runtime. See how it works, core components, and where legacy tools fall short.
- [MCP Tool Poisoning: How the Attack Works & How to Prevent It](https://agen.co/learning-center/mcp-tool-poisoning): MCP tool poisoning hides malicious instructions in tool descriptions that AI agents run as trusted commands. See how the attack works and how to shut it down.
- [RAG Security: Risks and Best Practices for Securing RAG](https://agen.co/learning-center/rag-security): RAG security treats retrieved context as untrusted input. See the top risks, from data poisoning to prompt injection, and a checklist to lock down the pipeline.
- [AI AppSec: A Threat-Surface Guide for AI Applications](https://agen.co/learning-center/ai-appsec): AI AppSec secures AI apps across the model, data, orchestration, and integration layers. Learn the threat surface, top risks, and the missing control plane.
- [AI Audit: A Complete Guide to Auditing AI Systems & Agents](https://agen.co/learning-center/ai-audit): What is an AI audit? What auditors examine, the process, audit trails, frameworks like NIST AI RMF, ISO 42001, and SOC 2, and how to get audit-ready for agents.
- [MCP Compliance: Governing MCP in Regulated Industries](https://agen.co/learning-center/mcp-compliance): MCP ships with no compliance layer. Learn how identity, audit logging, and data controls make Model Context Protocol meet SOC 2, GDPR, and the EU AI Act.
- [MCP Identity: Authentication & Authorization for MCP Servers](https://agen.co/learning-center/mcp-identity): MCP identity ties agents and tool calls to real principals. How OAuth 2.1, SSO, and least-privilege authorization work for MCP servers, clients, and agents.
- [AI Compliance Platform: What It Is and How to Choose One](https://agen.co/learning-center/ai-compliance-platform): An AI compliance platform monitors, enforces, and audits AI systems against the EU AI Act, ISO 42001, and NIST AI RMF. Learn what one does and how to choose.
- [MCP Call Filtering: Inspect and Block Risky Tool Calls](https://agen.co/learning-center/mcp-call-filtering): MCP call filtering inspects tool calls at the gateway, blocking risky ones by tool, parameter, or identity. Allowlist vs denylist and where enforcement belongs.
- [MITRE ATLAS Explained: Tactics, Techniques & AI Defenses](https://agen.co/learning-center/mitre-atlas): MITRE ATLAS catalogs real-world attacks on AI systems: tactics, techniques, and case studies. See how it maps to ATT&CK and how to defend your AI agents.
- [Prompt Injection: How Attacks Work & How to Stop Them](https://agen.co/learning-center/prompt-injection): Prompt injection is the top security risk for LLMs and AI agents. See how direct and indirect attacks work and get a defense-in-depth playbook to stop them.
- [What Is an MCP Registry? Enterprise Guide](https://agen.co/learning-center/mcp-registry): An MCP registry is a control plane, not a directory. Learn how MCP registries work, public vs. private, and how to govern, secure, and operate one at scale.
- [ISO/IEC 23894: The AI Risk Management Standard Explained](https://agen.co/learning-center/iso-iec-23894): ISO/IEC 23894 is the international standard for AI risk management. Learn what it covers, how it maps to NIST AI RMF and ISO 42001, and how to put it to work.
- [Multi Agent Systems in AI: How They Work & Secure Them](https://agen.co/learning-center/multi-agent-systems-ai): Multi agent systems in AI let many autonomous agents coordinate to solve problems one ag