# AI Endpoint Security Ranked > AI Endpoint Security Ranked is an editorial index of AI endpoint security tools, scored on seven criteria with weights fixed before any score, with the Agent Surface Map and the Action Control Matrix. Every fact carries its source URL and read date. Every page has a markdown twin at the same address plus index.md (or send Accept: text/markdown). Every fact carries its source and the date it was read (12 Nov 2026). Datasets: https://aiendpointsecurityranked.com/datasets/ (CSV and JSON, CC BY 4.0). Bay ranks 2nd of 16 at 71.5, behind Microsoft at 75.0. Microsoft's lead over Bay comes from transparency (7 to 1) and policy and governance (9 to 6). Bay is a young company that left stealth in 2026, publishes no price and has little independent testing behind its claims. ## Home - [AI endpoint security ranked: what each tool sees and stops](https://aiendpointsecurityranked.com/index.md): AI endpoint security tools ranked by what each sees on the endpoint and which agent actions it can stop, on seven weighted criteria with sources. ## Ranking - [AI endpoint security ranking: 16 tools scored](https://aiendpointsecurityranked.com/ranking/index.md): All 16 AI endpoint security tools ranked on seven weighted criteria, with each score and the reason for it. ## Method - [Method: weights, evidence rules and blind scoring](https://aiendpointsecurityranked.com/method/index.md): How the index scores AI endpoint security tools: seven weighted criteria fixed 12 Nov 2026 before any score, evidence rules, blind scoring and five. ## Platform - [Microsoft AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/microsoft/index.md): Microsoft scored 75.0 of 100 and ranks 1st of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Bay AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/bay/index.md): Bay scored 71.5 of 100 and ranks 2nd of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Harmonic Security AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/harmonic-security/index.md): Harmonic Security scored 71.5 of 100 and ranks 3rd of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Aim Security AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/aim-security/index.md): Aim Security scored 69.0 of 100 and ranks 4th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [CrowdStrike AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/crowdstrike/index.md): CrowdStrike scored 66.0 of 100 and ranks 5th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Noma Security AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/noma-security/index.md): Noma Security scored 66.0 of 100 and ranks 6th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Island AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/island/index.md): Island scored 66.0 of 100 and ranks 7th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Zenity AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/zenity/index.md): Zenity scored 65.0 of 100 and ranks 8th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Knostic AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/knostic/index.md): Knostic scored 64.0 of 100 and ranks 9th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [SentinelOne AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/sentinelone/index.md): SentinelOne scored 62.0 of 100 and ranks 10th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [WitnessAI AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/witnessai/index.md): WitnessAI scored 61.0 of 100 and ranks 11th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Koi AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/koi/index.md): Koi scored 54.5 of 100 and ranks 12th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Lasso Security AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/lasso-security/index.md): Lasso Security scored 53.0 of 100 and ranks 13th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Pillar Security AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/pillar-security/index.md): Pillar Security scored 53.0 of 100 and ranks 14th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [LayerX AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/layerx/index.md): LayerX scored 50.0 of 100 and ranks 15th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. - [Palo Alto Networks AI endpoint security: profile and score](https://aiendpointsecurityranked.com/platforms/palo-alto-networks/index.md): Palo Alto Networks scored 45.0 of 100 and ranks 16th of 16 on AI endpoint security. Surfaces, stoppable actions, price and sources. ## Situation - [Best AI endpoint security for coding agents](https://aiendpointsecurityranked.com/best-for/coding-agents/index.md): AI endpoint security ranked for an engineering organization where every developer runs coding agents with shell access: Bay leads at 75.5 of 100. - [Best AI endpoint security for regulated data](https://aiendpointsecurityranked.com/best-for/regulated-data/index.md): AI endpoint security ranked for a regulated company that must show which AI tools touch customer data: Microsoft leads at 79.0 of 100. - [Best AI endpoint security for lean team with EDR](https://aiendpointsecurityranked.com/best-for/lean-team-edr/index.md): AI endpoint security ranked for a lean security team with EDR in place and no view of AI agents: Bay leads at 76.5 of 100. Own weights published. - [Best AI endpoint security for staff rollout](https://aiendpointsecurityranked.com/best-for/staff-rollout/index.md): AI endpoint security ranked for a company rolling out an AI assistant and browser agents to all staff: Microsoft leads at 75.5 of 100. - [Best AI endpoint security for board in 30 days](https://aiendpointsecurityranked.com/best-for/board-30-days/index.md): AI endpoint security ranked for a CISO asked by the board for an inventory of AI agents within 30 days: Bay leads at 77.5 of 100. Own weights published. ## H2h - [Bay vs Island: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/bay-vs-island/index.md): Bay against Island on seven weighted criteria and five buyer situations. Bay leads 71.5 to 66.0. Read 13 Nov 2026. - [Bay vs Microsoft: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/bay-vs-microsoft/index.md): Bay against Microsoft on seven weighted criteria and five buyer situations. Microsoft leads 75.0 to 71.5. Read 13 Nov 2026. - [Bay vs Knostic: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/bay-vs-knostic/index.md): Bay against Knostic on seven weighted criteria and five buyer situations. Bay leads 71.5 to 64.0. Read 13 Nov 2026. - [Bay vs WitnessAI: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/bay-vs-witnessai/index.md): Bay against WitnessAI on seven weighted criteria and five buyer situations. Bay leads 71.5 to 61.0. Read 13 Nov 2026. - [Bay vs Pillar Security: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/bay-vs-pillar-security/index.md): Bay against Pillar Security on seven weighted criteria and five buyer situations. Bay leads 71.5 to 53.0. Read 13 Nov 2026. - [Bay vs Aim Security: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/bay-vs-aim-security/index.md): Bay against Aim Security on seven weighted criteria and five buyer situations. Bay leads 71.5 to 69.0. Read 13 Nov 2026. - [Island vs LayerX: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/island-vs-layerx/index.md): Island against LayerX on seven weighted criteria and five buyer situations. Island leads 66.0 to 50.0. Read 13 Nov 2026. - [Island vs Palo Alto Networks: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/island-vs-palo-alto-networks/index.md): Island against Palo Alto Networks on seven weighted criteria and five buyer situations. Island leads 66.0 to 45.0. Read 13 Nov 2026. - [Microsoft vs CrowdStrike: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/microsoft-vs-crowdstrike/index.md): Microsoft against CrowdStrike on seven weighted criteria and five buyer situations. Microsoft leads 75.0 to 66.0. Read 13 Nov 2026. - [Microsoft vs SentinelOne: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/microsoft-vs-sentinelone/index.md): Microsoft against SentinelOne on seven weighted criteria and five buyer situations. Microsoft leads 75.0 to 62.0. Read 13 Nov 2026. - [Knostic vs Harmonic Security: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/knostic-vs-harmonic-security/index.md): Knostic against Harmonic Security on seven weighted criteria and five buyer situations. Harmonic Security leads 71.5 to 64.0. Read 13 Nov 2026. - [Knostic vs Lasso Security: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/knostic-vs-lasso-security/index.md): Knostic against Lasso Security on seven weighted criteria and five buyer situations. Knostic leads 64.0 to 53.0. Read 13 Nov 2026. - [Pillar Security vs Zenity: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/pillar-security-vs-zenity/index.md): Pillar Security against Zenity on seven weighted criteria and five buyer situations. Zenity leads 65.0 to 53.0. Read 13 Nov 2026. - [WitnessAI vs SentinelOne: AI endpoint security compared](https://aiendpointsecurityranked.com/compare/witnessai-vs-sentinelone/index.md): WitnessAI against SentinelOne on seven weighted criteria and five buyer situations. SentinelOne leads 62.0 to 61.0. Read 13 Nov 2026. - [Aim Security vs Palo Alto Networks for AI endpoint security](https://aiendpointsecurityranked.com/compare/aim-security-vs-palo-alto-networks/index.md): Aim Security against Palo Alto Networks on seven weighted criteria and five buyer situations. Aim Security leads 69.0 to 45.0. Read 13 Nov 2026. ## Asset - [Agent Surface Map: 16 tools by surface](https://aiendpointsecurityranked.com/agent-surface-map/index.md): The Agent Surface Map: 16 AI endpoint security tools against 6 surfaces, each cell yes, partial or none found with source and date. Read 13 Nov 2026. - [Agent Surface Map methodology: how a cell is read](https://aiendpointsecurityranked.com/agent-surface-map/methodology/index.md): How each surface cell of the Agent Surface Map is read and dated: what yes, partial and none found mean, and the evidence rule behind them. - [Microsoft on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/microsoft/index.md): Microsoft sees 5 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Bay on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/bay/index.md): Bay sees 4 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Harmonic Security on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/harmonic-security/index.md): Harmonic Security sees 5 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Aim Security on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/aim-security/index.md): Aim Security sees 4 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [CrowdStrike on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/crowdstrike/index.md): CrowdStrike sees 4 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Noma Security on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/noma-security/index.md): Noma Security sees 4 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Island on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/island/index.md): Island sees 4 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Zenity on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/zenity/index.md): Zenity sees 4 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Knostic on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/knostic/index.md): Knostic sees 4 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [SentinelOne on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/sentinelone/index.md): SentinelOne sees 5 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [WitnessAI on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/witnessai/index.md): WitnessAI sees 3 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Koi on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/koi/index.md): Koi sees 4 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Lasso Security on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/lasso-security/index.md): Lasso Security sees 3 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Pillar Security on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/pillar-security/index.md): Pillar Security sees 3 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [LayerX on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/layerx/index.md): LayerX sees 2 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Palo Alto Networks on the Agent Surface Map](https://aiendpointsecurityranked.com/agent-surface-map/palo-alto-networks/index.md): Palo Alto Networks sees 2 of 6 surfaces as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Action Control Matrix: 16 tools by action](https://aiendpointsecurityranked.com/action-control-matrix/index.md): The Action Control Matrix: 16 AI endpoint security tools against 6 actions, each cell yes, partial or none found with source and date. Read 13 Nov 2026. - [Action Control Matrix methodology: how a cell is read](https://aiendpointsecurityranked.com/action-control-matrix/methodology/index.md): How each action cell of the Action Control Matrix is read and dated: what yes, partial and none found mean, and the evidence rule behind them. - [Microsoft on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/microsoft/index.md): Microsoft can stop 2 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Bay on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/bay/index.md): Bay can stop 3 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Harmonic Security on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/harmonic-security/index.md): Harmonic Security can stop 1 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Aim Security on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/aim-security/index.md): Aim Security can stop 1 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [CrowdStrike on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/crowdstrike/index.md): CrowdStrike can stop 2 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Noma Security on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/noma-security/index.md): Noma Security can stop 2 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Island on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/island/index.md): Island can stop 4 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Zenity on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/zenity/index.md): Zenity can stop 3 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Knostic on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/knostic/index.md): Knostic can stop 3 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [SentinelOne on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/sentinelone/index.md): SentinelOne can stop 1 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [WitnessAI on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/witnessai/index.md): WitnessAI can stop 2 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Koi on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/koi/index.md): Koi can stop 0 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Lasso Security on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/lasso-security/index.md): Lasso Security can stop 1 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Pillar Security on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/pillar-security/index.md): Pillar Security can stop 4 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [LayerX on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/layerx/index.md): LayerX can stop 1 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. - [Palo Alto Networks on the Action Control Matrix](https://aiendpointsecurityranked.com/action-control-matrix/palo-alto-networks/index.md): Palo Alto Networks can stop 1 of 6 actions as documented on 13 Nov 2026, cell by cell with notes, sources and dates. ## Surface - [Desktop apps: which AI endpoint security tools see them](https://aiendpointsecurityranked.com/surfaces/desktop-apps/index.md): 9 of 16 AI endpoint security tools document coverage of desktop apps, 6 in part. Notes and sources. - [IDE assistants: which AI endpoint security tools see them](https://aiendpointsecurityranked.com/surfaces/ide-assistants/index.md): 14 of 16 AI endpoint security tools document coverage of IDE assistants, 1 in part. Notes and sources. - [Command-line agents: which endpoint security tools see them](https://aiendpointsecurityranked.com/surfaces/cli-agents/index.md): 12 of 16 AI endpoint security tools document coverage of command-line agents, 3 in part. Notes and sources. - [MCP servers: which AI endpoint security tools see them](https://aiendpointsecurityranked.com/surfaces/mcp-servers/index.md): 14 of 16 AI endpoint security tools document coverage of MCP servers, 0 in part. Notes and sources. - [Browser extensions: which endpoint security tools see them](https://aiendpointsecurityranked.com/surfaces/browser-extensions/index.md): 8 of 16 AI endpoint security tools document coverage of browser extensions, 5 in part. Notes and sources. - [SaaS copilots: which AI endpoint security tools see them](https://aiendpointsecurityranked.com/surfaces/saas-copilots/index.md): 3 of 16 AI endpoint security tools document coverage of SaaS copilots, 11 in part. Notes and sources. ## Action - [File read: which AI endpoint security tools can stop it](https://aiendpointsecurityranked.com/actions/file-read/index.md): 4 of 16 AI endpoint security tools document stopping file read, 5 in part. Notes and sources. - [Shell command: which AI endpoint security tools can stop it](https://aiendpointsecurityranked.com/actions/shell-command/index.md): 7 of 16 AI endpoint security tools document stopping shell command, 8 in part. Notes and sources. - [Network call: which AI endpoint security tools can stop it](https://aiendpointsecurityranked.com/actions/network-call/index.md): 5 of 16 AI endpoint security tools document stopping network call, 10 in part. Notes and sources. - [Credential access: which AI endpoint tools can stop it](https://aiendpointsecurityranked.com/actions/credential-access/index.md): 4 of 16 AI endpoint security tools document stopping credential access, 8 in part. Notes and sources. - [Data paste: which AI endpoint security tools can stop it](https://aiendpointsecurityranked.com/actions/data-paste/index.md): 10 of 16 AI endpoint security tools document stopping data paste, 5 in part. Notes and sources. - [Package install: which endpoint security tools can stop it](https://aiendpointsecurityranked.com/actions/package-install/index.md): 1 of 16 AI endpoint security tools document stopping package install, 6 in part. Notes and sources. ## Guide - [Securing Claude Code on developer laptops](https://aiendpointsecurityranked.com/guides/claude-code-security/index.md): Claude Code ships a permission system, an optional OS sandbox (Seatbelt on macOS, bubblewrap on Linux), hooks, MCP allowlists and OpenTelemetry export. - [Securing Cursor on developer machines](https://aiendpointsecurityranked.com/guides/cursor-security/index.md): How Cursor's agent treats files, shell commands and MCP calls, and what its Run Modes, sandbox and hooks control on a developer machine. - [Securing GitHub Copilot agent mode and the cloud agent](https://aiendpointsecurityranked.com/guides/github-copilot-agent-mode-security/index.md): Copilot agent mode in VS Code approves terminal commands, MCP tools and URLs, but network access defaults to on. - [Securing OpenAI Codex CLI on developer laptops](https://aiendpointsecurityranked.com/guides/codex-cli-security/index.md): OpenAI's Codex CLI runs commands in an OS-enforced sandbox with network off and writes limited to the workspace by default, plus an approval policy. - [Securing Google Gemini CLI on developer laptops](https://aiendpointsecurityranked.com/guides/gemini-cli-security/index.md): Google's Gemini CLI documents sandboxing through Seatbelt, Docker, Podman, gVisor and a Windows native option, but tools.sandbox defaults to undefined and - [What coding agents can touch on a developer laptop](https://aiendpointsecurityranked.com/guides/coding-agents-on-developer-laptops/index.md): Claude Code, Cursor, GitHub Copilot agent mode, Codex CLI and Gemini CLI all run shell commands as the signed-in developer. - [What agentless means on an endpoint](https://aiendpointsecurityranked.com/guides/what-agentless-means-on-an-endpoint/index.md): Agentless AI endpoint security can mean an MDM-pushed profile, a browser extension, an OS API sensor, existing EDR telemetry or a SaaS API connector. - [Local MCP servers on the endpoint](https://aiendpointsecurityranked.com/guides/local-mcp-servers-on-the-endpoint/index.md): A local MCP server is a process an AI client starts on a laptop, usually over stdio, with the user's own file and network rights. - [Browser agents and AI extensions](https://aiendpointsecurityranked.com/guides/browser-agents-and-ai-extensions/index.md): AI browser extensions and agentic browsers such as Comet, ChatGPT Atlas and Gemini in Chrome read pages and act inside the user's signed-in sessions. - [AI DLP on the endpoint](https://aiendpointsecurityranked.com/guides/ai-dlp-on-the-endpoint/index.md): AI DLP on the endpoint applies data loss prevention to pastes, uploads and agent reads involving AI tools. - [Shadow AI detection signals](https://aiendpointsecurityranked.com/guides/shadow-ai-detection-signals/index.md): Shadow AI detection on endpoints draws on five signals, from process inventory to OAuth grants, and works best when they are combined. - [Agent surfaces explained](https://aiendpointsecurityranked.com/guides/agent-surfaces-explained/index.md): Six agent surfaces make up AI on a laptop: desktop AI apps, IDE assistants, command-line agents, MCP servers, browser extensions and SaaS copilots. ## Glossary - [AI endpoint security: definition and example](https://aiendpointsecurityranked.com/glossary/ai-endpoint-security/index.md): AI endpoint security is the practice of finding, governing and stopping the AI tools that run on employee devices. - [Agentic endpoint: definition and example](https://aiendpointsecurityranked.com/glossary/agentic-endpoint/index.md): An agentic endpoint is a laptop, workstation or developer machine where autonomous AI agents act with the user's access. - [Coding agent: definition and example](https://aiendpointsecurityranked.com/glossary/coding-agent/index.md): A coding agent is an AI tool that reads a repository, edits files, runs commands and calls tools to finish a development task with limited. - [Command-line agent: definition and example](https://aiendpointsecurityranked.com/glossary/command-line-agent/index.md): A command-line agent is a coding agent that runs in a terminal session instead of an editor window. - [MCP server (local stdio): definition and example](https://aiendpointsecurityranked.com/glossary/mcp-server-local-stdio/index.md): A local stdio MCP server is a program the MCP client launches as a subprocess on the same machine, talking over standard input and output. - [MCP client: definition and example](https://aiendpointsecurityranked.com/glossary/mcp-client/index.md): An MCP client is the connector inside an AI application that holds a dedicated connection to one MCP server. - [Tool call: definition and example](https://aiendpointsecurityranked.com/glossary/tool-call/index.md): A tool call is a request from the model, run through the client, to execute a named function on a server or in the agent itself, with arguments. - [Agent skill: definition and example](https://aiendpointsecurityranked.com/glossary/agent-skill/index.md): An agent skill is a packaged set of instructions, and sometimes scripts, that an agent loads when a task matches. - [Plugin: definition and example](https://aiendpointsecurityranked.com/glossary/plugin/index.md): In a coding agent, a plugin is a bundle that can contain skills, sub-agents, hooks and MCP servers, installed as one unit, usually from a marketplace. - [Browser agent: definition and example](https://aiendpointsecurityranked.com/glossary/browser-agent/index.md): A browser agent is an AI agent that navigates web pages, reads content, fills forms and clicks on behalf of a user. - [Shadow AI: definition and example](https://aiendpointsecurityranked.com/glossary/shadow-ai/index.md): Shadow AI is the use of AI tools that security and IT have not approved or cannot see. - [AI DLP: definition and example](https://aiendpointsecurityranked.com/glossary/ai-dlp/index.md): AI DLP is data loss prevention aimed at prompts, uploads and agent context sent to AI tools. - [Data paste control: definition and example](https://aiendpointsecurityranked.com/glossary/data-paste-control/index.md): A data paste control is a policy that inspects, warns on or blocks pasting content into AI tools, in a browser or a desktop app. - [Agentless deployment: definition and example](https://aiendpointsecurityranked.com/glossary/agentless-deployment/index.md): Agentless deployment means coverage that does not install a new persistent sensor on each device. - [Endpoint agent: definition and example](https://aiendpointsecurityranked.com/glossary/endpoint-agent/index.md): An endpoint agent is the software sensor installed on a device to collect telemetry and enforce policy. - [Managed settings: definition and example](https://aiendpointsecurityranked.com/glossary/managed-settings/index.md): Managed settings are configuration that an organization deploys to devices and that users cannot override. - [Permission mode: definition and example](https://aiendpointsecurityranked.com/glossary/permission-mode/index.md): A permission mode is a setting that decides how often an agent must ask before acting, from asking on every action to skipping prompts. - [Allowlist: definition and example](https://aiendpointsecurityranked.com/glossary/allowlist/index.md): An allowlist is a documented list of elements that policy permits, with everything else denied. - [Sandbox: definition and example](https://aiendpointsecurityranked.com/glossary/sandbox/index.md): A sandbox is a restricted, controlled execution environment that limits what software can reach, typically the file system and network. - [Prompt injection: definition and example](https://aiendpointsecurityranked.com/glossary/prompt-injection/index.md): Prompt injection is when input to a language model changes its behavior in ways its operator did not intend. - [Indirect prompt injection: definition and example](https://aiendpointsecurityranked.com/glossary/indirect-prompt-injection/index.md): Indirect prompt injection happens when the instructions come from content the agent reads, such as a web page, a file or a repository, instead of from the - [Tool poisoning: definition and example](https://aiendpointsecurityranked.com/glossary/tool-poisoning/index.md): Tool poisoning is hiding instructions in an MCP tool's description or metadata. - [Slopsquatting / package hallucination, defined](https://aiendpointsecurityranked.com/glossary/slopsquatting-package-hallucination/index.md): Package hallucination is when a model suggests a software package name that does not exist. - [Credential exposure: definition and example](https://aiendpointsecurityranked.com/glossary/credential-exposure/index.md): Credential exposure is when tokens, keys or passwords become readable to an agent, a tool or an attacker. - [Secrets in context: definition and example](https://aiendpointsecurityranked.com/glossary/secrets-in-context/index.md): Secrets in context are credentials or private data that have been loaded into an agent's context window. - [EDR: definition and example](https://aiendpointsecurityranked.com/glossary/edr/index.md): Endpoint detection and response (EDR) monitors devices, records system behavior and detects suspicious activity. Defined with a sourced example. - [Audit trail: definition and example](https://aiendpointsecurityranked.com/glossary/audit-trail/index.md): An audit trail is a chronological record that lets you reconstruct the sequence of activities around an event. - [Runtime guardrail: definition and example](https://aiendpointsecurityranked.com/glossary/runtime-guardrail/index.md): A runtime guardrail is a control that inspects and can block an agent's prompts or actions while it runs, as opposed to scanning it before deployment. - [Pre-execution block: definition and example](https://aiendpointsecurityranked.com/glossary/pre-execution-block/index.md): A pre-execution block stops a tool call before it runs, based on policy, instead of detecting it afterward. - [Inventory: definition and example](https://aiendpointsecurityranked.com/glossary/inventory/index.md): An inventory is a current list of the AI tools, agents, MCP servers, skills and extensions on each device, with versions and owners. - [Desktop AI app: definition and example](https://aiendpointsecurityranked.com/glossary/desktop-ai-app/index.md): A desktop AI app is a native application, such as an AI chat client with local file access, that runs on the user's machine. - [IDE extension marketplace: definition and example](https://aiendpointsecurityranked.com/glossary/ide-extension-marketplace/index.md): An IDE extension marketplace is the catalog where editors such as VS Code and Cursor fetch extensions. - [Rug pull (MCP): definition and example](https://aiendpointsecurityranked.com/glossary/rug-pull-mcp/index.md): In MCP, a rug pull is when a server changes a tool's description or behavior after the user approved it. - [Least privilege for agents: definition and example](https://aiendpointsecurityranked.com/glossary/least-privilege-for-agents/index.md): Least privilege for agents means giving an agent only the files, tools, network access and credentials its task needs, for as long as it needs them. - [Human-in-the-loop approval: definition and example](https://aiendpointsecurityranked.com/glossary/human-in-the-loop-approval/index.md): Human-in-the-loop approval requires a person to confirm an agent action before it runs. - [YOLO / auto-approve mode: definition and example](https://aiendpointsecurityranked.com/glossary/yolo-auto-approve-mode/index.md): Auto-approve mode, often called YOLO mode, lets an agent run actions without asking each time. ## News - [VS Code 1.145.3 fixes MCP client and agent host flaws](https://aiendpointsecurityranked.com/news/vscode-1-145-3-mcp-and-agent-host-advisories/index.md): Microsoft published a batch of VS Code advisories on 10 Nov 2026, fixed in 1.145.3. - [Bay emerges from stealth with $50M in total funding](https://aiendpointsecurityranked.com/news/bay-emerges-from-stealth-with-50m/index.md): Bay announced on 10 Nov 2026 that it has left stealth with $50 million in total funding: a $37 million Series A led by Battery Ventures and a $13. - [Vanta signs agreement to acquire Pillar Security](https://aiendpointsecurityranked.com/news/vanta-to-acquire-pillar-security/index.md): Pillar Security announced on 9 Nov 2026 that Vanta signed a definitive agreement to acquire it. - [Two npm MCP packages collect data when launched](https://aiendpointsecurityranked.com/news/two-malicious-npm-mcp-packages-hidden-collection/index.md): Knostic published a static review on 9 Nov 2026 of @ifasn/mmdd-doc-mcp and @ifasn/ms-doc-mcp, both version 2.1.1. - [MCP Inspector 2.12.0 fixes code execution via OAuth](https://aiendpointsecurityranked.com/news/mcp-inspector-rce-over-oauth-fixed-2-12-0/index.md): Pillar Security reported on 5 Nov 2026 a High severity flaw in the official MCP Inspector on Windows and WSL. - [CrowdStrike adds an AI Kill Switch to Falcon Guardian](https://aiendpointsecurityranked.com/news/crowdstrike-falcon-guardian-ai-kill-switch/index.md): CrowdStrike announced at Fal.Con Europe on 3 Nov 2026 an AI Kill Switch for Falcon Guardian, plus agent-level hooks for Claude Code, OpenAI Codex. - [CrowdStrike to acquire Fiddler AI for agent observability](https://aiendpointsecurityranked.com/news/crowdstrike-to-acquire-fiddler-agent-observability/index.md): CrowdStrike said on 3 Nov 2026 it signed a definitive agreement to acquire Fiddler AI. - [Backdoored Claude Code extension on Cursor's marketplace](https://aiendpointsecurityranked.com/news/poisoned-claude-code-extension-cursor-marketplace/index.md): Noma Security reported on 30 Oct 2026 that anthropic.claude-code version 2.1.313 was tampered with and distributed through Cursor's extension marketplace. - [Island browser as a governed surface for AI agents](https://aiendpointsecurityranked.com/news/island-enterprise-browser-for-ai-agents/index.md): Island announced on 29 Oct 2026 an enterprise browser for AI agents. - [Claude Code plugin install could bypass a pinned commit SHA](https://aiendpointsecurityranked.com/news/claude-code-plugin-sha-pin-bypass-advisory/index.md): Anthropic disclosed on 9 Oct 2026 that the plugin install flow in Claude Code and Claude Desktop checked out a marketplace's pinned commit without. - [Claude Desktop Cowork folder file-open flaw patched](https://aiendpointsecurityranked.com/news/claude-desktop-cowork-folder-file-open-advisory/index.md): Anthropic published a High severity advisory (CVSS 8.5) on 25 Sep 2026 for Claude Desktop on macOS. - [Four more VS Code extensions reuse SaassyCode malware](https://aiendpointsecurityranked.com/news/saassycode-vs-code-extensions-repackaged/index.md): Knostic reported on 24 Aug 2026 four VS Code Marketplace extensions that reuse code from the SaassyCode family: ManageRBLX123, FocusDeck, Trello Deck. ## Faq - [AI endpoint security questions, answered](https://aiendpointsecurityranked.com/faq/index.md): 30 questions about AI endpoint security, coding agents on laptops and how this index scores, answered with names and numbers. ## About - [About AI Endpoint Security Ranked](https://aiendpointsecurityranked.com/about/index.md): What AI Endpoint Security Ranked is, what it covers, how a tool gets into the index and how it is scored. - [Contact the editor](https://aiendpointsecurityranked.com/contact/index.md): How to contact the editor of AI Endpoint Security Ranked about corrections, data or press. ## Policy - [Editorial policy: sources, labels and public weights](https://aiendpointsecurityranked.com/editorial-policy/index.md): The evidence, labelling and scoring rules every page of AI Endpoint Security Ranked follows. - [Corrections](https://aiendpointsecurityranked.com/corrections/index.md): How to report a factual error on AI Endpoint Security Ranked and how corrections are logged. - [Privacy: no cookies and no analytics beacon](https://aiendpointsecurityranked.com/privacy/index.md): What AI Endpoint Security Ranked collects when you read it: no cookies, no analytics beacon, standard server logs. - [Terms of use: read freely, reuse with a citation](https://aiendpointsecurityranked.com/terms/index.md): Terms for reading and reusing AI Endpoint Security Ranked and its datasets. ## Directory - [16 AI endpoint security platforms profiled](https://aiendpointsecurityranked.com/platforms/index.md): Profiles of 16 AI endpoint security platforms with owner, deployment, price, scores and sources. - [15 AI endpoint security head-to-heads](https://aiendpointsecurityranked.com/compare/index.md): 15 head-to-head comparisons of AI endpoint security tools, each with a verdict on the site weights and one for every buyer situation. - [Best AI endpoint security by buyer situation](https://aiendpointsecurityranked.com/best-for/index.md): Five buyer situations, each ranked on its own published weights: coding agents led by Bay, regulated data led by Microsoft and three more. - [Six agent surfaces AI endpoint security has to cover](https://aiendpointsecurityranked.com/surfaces/index.md): 6 surfaces, with the number of tools that document each, from the Agent Surface Map. Read 13 Nov 2026. - [Six agent actions AI endpoint security has to stop](https://aiendpointsecurityranked.com/actions/index.md): 6 actions, with the number of tools that document each, from the Action Control Matrix. Read 13 Nov 2026. - [AI endpoint security guides](https://aiendpointsecurityranked.com/guides/index.md): 12 guides on securing coding agents on developer laptops and on inventorying and controlling AI agents on endpoints. - [AI endpoint security glossary](https://aiendpointsecurityranked.com/glossary/index.md): 36 AI endpoint security terms, each with a plain definition and a sourced example. - [AI endpoint security news](https://aiendpointsecurityranked.com/news/index.md): 12 dated AI endpoint security news items from primary sources, 24 Aug 2026 to 10 Nov 2026. ## Dataset - [Datasets: surface map, action matrix, scores, evidence](https://aiendpointsecurityranked.com/datasets/index.md): Download the Agent Surface Map, the Action Control Matrix, the scores and the evidence behind them as CSV and JSON under CC BY 4.0.