Skip to main content

Key Concepts

A quick glossary of the terms you'll see across the Endpoints surface and these docs. If you've read How it works, this formalizes the vocabulary.

Device & identity​

  • Managed device — a Mac or Windows machine running the Waxell agent, whether deployed by MDM or installed manually.
  • Device owner — the person who normally uses the device. Set on the device's page or filled in from your MDM; it drives team policy and user blocks, and attributes captures to a person.
  • Enrollment — the first-run handshake where a device registers with your workspace and receives its own device identity.
  • Device → agent map — on the Devices tab you map each device's AI apps to a governed agent, so captured activity and on-device tools resolve to that agent for policy and reporting.

AI apps​

  • AI app — any application the on-device scanner identifies as talking to an AI provider: Claude Desktop, Cursor, Claude Code, GitHub Copilot, ChatGPT desktop, Perplexity, browser assistants, and more.
  • App type — the class of an AI app (e.g. a CLI coding agent vs. a desktop chat app vs. a browser). Guards can be set per app type, not just per app.
  • Governance level — how much of an app's activity Waxell can observe, from Waxell's catalog: full (the actual request and response), transcript (the rendered conversation), metadata (who, where, when — no content) or none. Shown per app in Inventory and on the Guard page.

Capture​

  • Browser capture — AI conversations read in the page by the browser extension: prompt, reply, model, account type. On by default, redacted on the device before upload. Shown in Captures and Device Activity.
  • Built-in sites — ChatGPT, Claude, Gemini, Microsoft Copilot and Perplexity. Add others with Browser extra hosts.
  • Account type — whether a browser AI send came from a corporate, personal or unknown account.
  • Held — Browser capture is off for the device: captures stay on it instead of uploading.
  • Redacted — the capture had secrets removed on the device before upload.
  • Metadata flow — which app reached which AI host, when, as which user. No content. Shown in Activity → Flow.
  • Site adapter — the per-site rules that tell the extension how to read a given AI site. Maintained by Waxell, overridable per workspace.

Hooks​

  • Hook — a check a coding agent (Claude Code, Codex, Cursor…) runs before it acts. Waxell's hook answers allow, ask or deny from your Guard.
  • Managed hooks — hooks registered in the tool's administrator-owned configuration, which the user cannot override.
  • Named rule — an entry in the Guard catalog with a plain-English name (e.g. "Download-and-run") that stands for a command or file pattern.

Guard & policy​

  • Guard — the policy every device enforces: what is captured, allowed, asked about or blocked.
  • Guard cascade — a layered policy resolved in precedence order Global → App type → Team → Device → Agent group → Agent. More-specific layers override less-specific ones field by field. See Guard cascade overview.
  • Layer / override — a partial config set at one scope of the cascade. You only specify the fields you want to change; everything else inherits from the layer above.
  • Resolve — the Guard page's resolver: pick an app type / team / device / agent and see the effective config and which layer each field came from.

Detective vs. preventive​

  • Preventive — stopped before it happens: a hook denying a command, the extension refusing a prompt, a blocked site or app.
  • Detective — recorded after the fact: browser captures, transcripts, the ChatGPT desktop app. The Decisions page labels which is which.
  • Containment — device-level blocks and quarantine, which cuts an app off after a serious violation. See Containment.

Where things are​

Everything is under Endpoints in your workspace: Overview, Activity, Captures, Alerts, Inventory, Guard, Decisions, Devices, Health, Site Adapters and Setup. See How it works.

Next: deploy to a fleet or a single machine, then set up Guard.