The cross-harness, self-improving memory layer for AI agents. Join our community: https://discord.com/invite/zdNChS2fBu
Go agent-ai-cliagent-harnessesagent-securitybrowser-extensionciclaude-codecloudcodexcursordetection-engineeringjevknowledge-basememorymobile-device-managementobservabilitysecuritysecurity-information-and-event-managementsecurity-toolstelemetrytraces
fix
Justin D'Souza05b8a96da8c6
diff --git a/README.md b/README.mdindex 03885305..d5c99271 100644--- a/README.md+++ b/README.md@@ -13,8 +13,26 @@ Beacon is visibility-first. The current public build focuses on observing localagent runtime activity, normalizing it into endpoint events, and leavingforwarding to existing localfile/Wazuh or customer-managed pipelines.+## Product Vision++Beacon is built around a simple thesis: as AI agents move from answering+questions to taking actions on laptops and workstations, the endpoint becomes+the place where intent, context, permissions, tools, credentials, and changes+come together.++The long-term vision is to make local agent activity observable,+understandable, and eventually governable across the enterprise. Beacon starts+with the visibility layer: a common, local-first record of what agents were+asked to do, what context they accessed, which tools and commands they used,+what files they touched, what approvals were granted, and what changed on the+endpoint.++Read more in+[Introducing Beacon: Endpoint Telemetry for AI Agents](https://justindsouza.substack.com/p/introducing-beacon-endpoint-telemetry).+## Table Of Contents+- [Product Vision](#product-vision)- [What Beacon Does](#what-beacon-does)- [Privacy And Retention](#privacy-and-retention)- [What Beacon Does Not Do](#what-beacon-does-not-do)