Analysis updated 2026-05-18
Log prompts, commands, tool use, and file edits from AI coding assistants on developer laptops.
Deploy Beacon across a fleet of Macs using Jamf Pro or Fleet MDM.
Forward captured AI agent events to Wazuh or Splunk for centralized security monitoring.
| asymptote-labs/agent-beacon | saiyam1814/kiac | rockorager/comview | |
|---|---|---|---|
| Stars | 44 | 45 | 47 |
| Language | Go | Go | Go |
| Setup difficulty | hard | moderate | easy |
| Complexity | 4/5 | 3/5 | 2/5 |
| Audience | ops devops | ops devops | developer |
Figures from each repo's GitHub metadata at analysis time.
Fleet deployment requires macOS MDM tooling such as Jamf Pro or Fleet, and a signed, notarized package install.
Beacon is an open source monitoring tool that watches what AI coding assistants are doing on a developer's computer and produces a structured log of that activity for security teams to review. As AI agents like coding assistants start taking real actions such as running commands, reading and editing files, and making network requests, organizations face a new visibility gap: they can see what applications are installed but have no unified record of what the AI actually did during a session. Beacon runs quietly in the background on the endpoint, the developer's laptop or workstation, and intercepts telemetry from supported AI runtimes. It captures things like which prompts were sent, which tools the agent invoked, which commands it ran, which files it touched, and what approvals were granted or denied. All of this is written to a local log file in a structured format compatible with Wazuh, a popular open source security information platform. Teams that use Splunk, another security monitoring system, can forward the same events there as an optional step. The tool is deliberately local first: it does not send data to any external server during normal operation, which matters for organizations that handle sensitive code. A local only dashboard lets individuals inspect their own agent activity in a browser without sending anything off the machine. Security and IT teams can deploy Beacon across a fleet of developer Macs using MDM tools like Jamf Pro or Fleet. It is written in Go and currently supports macOS.
A local, open source telemetry tool that logs what AI coding agents do on a developer's machine for security teams.
Mainly Go. The stack also includes Go, Wazuh, Splunk HEC.
No license information is provided in the README.
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly ops devops.
This repo across BitVibe Labs
Verify against the repo before relying on details.