agent-beacon

by Asymptote-LabsVerified

Agent Beacon is the world's first open-source telemetry layer for AI agents wherever they run: locally, in CI, or in the cloud.

303
Stars
13
Forks
Go
Language
8/23/2026
Added
View on GitHubDownload ZIP

⚠️ Third-Party Software Notice

This skill is third-party open-source software developed and hosted independently on GitHub. SkillTip is an informational directory and does not control or maintain the underlying repository. Any security checks displayed are automated and limited in scope. Review the source code before installing.

Read the Terms of Service

Installation

Add to your Claude Code skills directory:

# Add to your Claude Code skills
git clone https://github.com/Asymptote-Labs/agent-beacon

Getting Started

Guides for using skills like agent-beacon.

Security Report

Verified

Last scanned: —

{
  "status": "PASSED",
  "issues": []
}

README.md

Beacon

Asymptote Lab's Agent Beacon

GitHub release Homebrew GitHub Workflow Status MIT license Docs Discord

Unified telemetry for AI agents, wherever they run.

Docs · Discord · Install · For Security & IT Teams · Dashboard · Commands

What is Agent Beacon

Agent Beacon is the world's first open-source telemetry layer for AI agents wherever they run: locally, in CI, or in the cloud.

The problem is that AI agent activity is fragmented across runtimes, leaving teams without a consistent way to get visibility into what agents are doing. Agent Beacon solves this by extending the OpenTelemetry GenAI standard and normalizing runtime events into a unified data model.

Beacon is built to be easy to deploy for Security and IT teams through MDM deployment, CI workflows, and cloud-agent setup paths, and to emit agent harness telemetry logs to all the major enterprise-grade SIEMs.

Learn more in the Agent Beacon Documentation.

High-Level Architecture

Beacon keeps endpoint collection, processing, and inspection local by default, while extending the same normalized event model to CI and cloud-agent telemetry paths under customer control.

Beacon endpoint architecture

  • Agent runtime layer: Hooks, OpenTelemetry sources, CI wrappers, and SDKs capture supported activity from AI agent harnesses wherever they run.
  • Beacon endpoint layer: Local processing normalizes events, applies retention and redaction settings, and writes durable endpoint telemetry.
  • Output layer: Teams inspect events in the local dashboard, retain JSONL, or forward records into all the major enterprise-grade SIEMs.

Supported Surfaces

Beacon captures supported agent harness activity across local endpoints, CI jobs, and cloud-agent surfaces, then writes normalized events that teams can inspect in place or forward into customer-managed security pipelines.

Agent Runtimes

Agent Beacon supports the most popular enterprise agent harnesses across local, CI, and cloud surfaces.

Local Agents

Coding Agent Harnesses
Agent harnessCollection pathTelemetry coverage
Antigravity CLINative hooksPrompt, pre-tool, post-tool, stop, invocation, command, and file telemetry where Antigravity exposes hook payloads
Claude CodeLocal OTLP export plus optional hooksPrompt, command, tool, file, approval, API/model lifecycle, MCP connection, subagent, and session telemetry where emitted through OTLP or hooks
ClineManaged plugin hooksPrompts, task lifecycle/errors, tool lifecycle/results, commands with exit codes, file reads/edits with diffs, MCP activity, and task token usage/cost
Codex CLILocal OTLP logsSession, prompt, approval, and tool-result activity from Codex semantic logs
CursorNative hooksPrompt, tool, shell command, MCP-like, approval, and file edit telemetry
Devin CLINative hooksSession, prompt, pre-tool, post-tool, permission request, stop, session-end, approval, and file telemetry
Devin DesktopCascade/Windsurf hooksPrompt, command, MCP tool, file read, and file write telemetry where Desktop exposes Cascade hook payloads
Factory DroidOTLP HTTP plus optional hooksSession, prompt, write/edit/create tool use, stop, session-end, and available OTLP telemetry
Gemini CLIOpt-in local OTLPPrompts, tool calls, MCP activity, file operations, and approval-related events emitted through OTLP
GitHub Copilot CLIMDM-managed OTLP HTTPPrompt, session, tool, and approval-like activity emitted through Copilot CLI spans
Grok BuildNative hooksSession, prompt, pre-tool, post-tool, failed tool, stop, session-end, command, and file telemetry
OpenCodeManaged plugin hooksPrompts, assistant output/reasoning, model usage/cost, tool lifecycle/results, commands, file/web/MCP activity, approvals, and session errors
PiManaged extension hooksSession lifecycle, prompts, tool lifecycle/results, commands including operator ! commands, file reads/writes/edits with diffs, agent reasoning, and token usage/cost
Qwen CodeNative hooksSession, prompt, pre-tool, post-tool, failed tool, permission request/approval, subagent, stop, session-end, command, and file telemetry merged into Qwen's own settings.json
VS CodeCopilot Chat OTel plus optional preview hooksCopilot session, prompt, model, and tool activity through OTel; optional hooks for extra lifecycle and cross-agent detail
Knowledge Worker Agent Harnesses
Agent harnessCollection pathTelemetry coverage
Claude CoworkAdmin-configured OTLPPrompt, command, tool, and file telemetry when emitted through Claude Cowork OTLP
Hermes AgentShell hooksPrompt, observed tool, command, file, approval request and response, session lifecycle, and subagent stop telemetry
OpenClaw GatewayGateway-configured OTLP/HTTPOTLP logs, traces, and metrics from the Gateway diagnostics plugin

CI Agents

HarnessCollection pathTelemetry coverage
CI agent telemetryTemporary local collector through beacon ci exec or beacon ci start / beacon ci finishSupported agent prompt, tool, command, file, and run context where emitted during the job

Cloud Agents

Cloud surfaceCollection pathTelemetry coverage
AnthropicOpenLLMetry instrumentation through @asymptote/sdkSupported Anthropic model call spans, errors, and OpenTelemetry attributes
Claude Agent SDKQuery wrapper through Observe.wrapClaudeAgentQuery()Query root spans with Beacon-compatible prompt attributes
Claude Code Cloud AgentsCloud sandbox hooks with direct GCS or S3 uploadSession, prompt, tool, command, file, and lifecycle telemetry where Claude Code cloud hook payloads expose it
Cursor Cloud AgentsCloud sandbox hooks with direct GCS or S3 uploadFollow-up prompts, tool, shell command, file, subagent, and compaction telemetry after project hooks become active
Devin Cloud AgentsOrg-wide API poll via beacon cloud devin pull, with GCS uploadSession, prompt, agent message, status, pull request, and ACU usage telemetry the Devin sessions API exposes (message-level; the autonomous agent runs no in-sandbox hooks)
OpenAIOpenLLMetry instrumentation through @asymptote/sdkSupported OpenAI model call spans, errors, and OpenTelemetry attributes
Vercel AI SDKTracer handoff through experimental_telemetryAI SDK model call and tool spans where telemetry is enabled

Output Destinations

Agent Beacon writes endpoint telemetry to local JSONL by default and supports customer-controlled forwarding into common security information and event management (SIEM), log aggregation, and object storage destinations.

Security Information and Event Management (SIEM)

DestinationSupport path
CrowdStrike Falcon LogScale HECOptional endpoint forwarding with LogScale ingest tokens during install or repair
Microsoft SentinelAzure Monitor Agent and Data Collection Rule content pack over local JSONL
Rapid7 InsightIDRCustom Logs webhook content pack over local JSONL
Splunk HECOptional endpoint forwarding during install or repair
Sumo LogicHTTP Logs & Metrics Source content pack over local JSONL
WazuhLocalfile configuration and Beacon Wazuh content pack

Log Aggregation

DestinationSupport path
AWS CloudWatch LogsVector content pack over local JSONL using customer-managed AWS credentials
Customer-managed log pipelinesForwarding from local Beacon JSONL under customer control
DatadogDatadog Agent custom log collection over local JSONL
ElasticFilebeat or Elastic Agent content pack over local JSONL

Object Storage

DestinationSupport path
AWS S3Vector over endpoint JSONL, CI upload, or direct compressed snapshots from supported cloud agents
Google Cloud StorageVector and packaged macOS helpers over endpoint JSONL, CI upload, or direct compressed snapshots from supported cloud agents

Local

DestinationSupport path
Local JSONLDefault endpoint log and local dashboard source

MDM Deployment

Agent Beacon is designed for Security and IT teams to deploy and validate through standard MDM workflows.

Version tags publish a signed, notarized, and stapled Apple Silicon endpoint .pkg to GitHub Releases for MDM/manual download, .deb and .rpm packages for Linux on amd64 and arm64, and an x64 .msi for Windows. Homebrew and release archives remain available for CLI installs across supported macOS, Linux, and Windows architectures.

Installing a native package performs the system-mode install itself: it registers and starts the service, writes configuration to the platform's machine-wide location, and points the interactive user's agent runtimes at the local collector. See the Linux install guide or the Windows install guide.

PlatformPackageService managerNotes
macOSSigned, notarized .pkg (Apple Silicon)launchdHomebrew for single machines
Linux.deb / .rpm (amd64, arm64)systemdSupervised fallback without systemd
Windows.msi (x64)Service Control ManagerUnsigned for now; verify the published .sha256
MDM platformSupport path
FleetmacOS package and user-context deployment helpers
Jamf PromacOS package, policy scripts, validation, and Extension Attributes

The macOS package includes GCS forwarder helpers at /opt/beacon/jamf/claude/gcs/{install-forwarder.sh,run-forwarder.sh,repair-hooks-and-forwarder.sh}. They run bundled Vector as com.beacon.endpoint.gcs-forwarder, write runtime and inventory objects below one root prefix, and reference an externally delivered service-account JSON through GOOGLE_APPLICATION_CREDENTIALS.

Dashboard and Local Detection

Beacon includes a local, read-only dashboard for validating endpoint activity without a hosted backend. See the dashboard docs for overview, log search, and runtime JSONL inspection.

Beacon writes endpoint activity to runtime.jsonl and periodic Cursor/Claude Code configuration inventory metadata to the sibling inventory_state.jsonl. Local log storage and retention behavior are summarized in the local testing and logs docs.

For offline threat detection, beacon scan runs open threat rules over local telemetry with no network access. See the Threat Rules spec and generated rule field reference for rule format, CEL matching, fixtures, and supported event fields.

Start Here

Quickstart

See the Quickstart docs for the full setup paths.

First-run onboarding

The first time you run beacon endpoint install in a terminal, Beacon asks two questions — your email and whether this is work or personal use — and sends the answers to Asymptote once. Knowing who runs Beacon is how we decide which runtimes and integrations to build next.

Exactly what is sent, and nothing else:

FieldExample
Email you enteryou@company.com
Work, personal, or evaluatingwork
OS, architecture, OS versiondarwin, arm64, 15.5
Beacon version and install methodv0.0.31, homebrew
Names of agent runtimes on this machineclaude_code, cursor
A random install ID64871b2b…

Never sent: prompts, file contents, commands, telemetry events, repository names, or anything else Beacon captures. The endpoint agent itself stays local-only — this is one HTTP request at install time, not an ongoing channel.

It happens once per machine. The answer is recorded in ~/.beacon/profile.json, which survives uninstall so a reinstall does not ask again.

It never runs non-interactively. Package postinstall scripts, MDM deployments, --system installs, CI, --dry-run, and any piped or redirected stdin skip it silently. Unattended installs that still need it suppressed can set:

BEACON_ONBOARDING=0 beacon endpoint install

For a fleet rollout where you do want attribution but have no terminal, supply the answers up front:

BEACON_ONBOARDING_EMAIL=it@company.com BEACON_ONBOARDING_USAGE=work \
  beacon endpoint install

Inspect or clear the record at any time:

beacon endpoint onboarding          # show what was recorded
beacon endpoint onboarding --reset  # clear it

To have your record deleted, email the install ID shown by beacon endpoint onboarding to support@asymptotelabs.ai.

For Security & IT Teams

Start with the security and IT quickstart and managed deployment guidance for rollout, validation, retention, and SIEM forwarding. For vendor review, see the security review.

For Developers

Install the released Beacon CLI locally with Homebrew:

brew tap asymptote-labs/tap
brew install beacon
beacon version

Or build from source:

cd cli/beacon
make build

To verify a change against a real Claude Code session rather than only synthetic payloads, beacon-sandbox runs one in a disposable Linux sandbox and checks what Beacon actually captured:

cd beacon-sandbox
go run ./cmd/beacon-sandbox doctor
go run ./cmd/beacon-sandbox run --scenario s02-bash-command

See Verify Beacon In A Sandbox for setup, what it can verify, and its limitations.

For setup, deployment, integrations, and command details, see the Beacon CLI docs.

Star Growth

Beacon GitHub star growth

License

MIT

Frequently Asked Questions

What is agent-beacon?

agent-beacon is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by Asymptote-Labs. Agent Beacon is the world's first open-source telemetry layer for AI agents wherever they run: locally, in CI, or in the cloud. It has 303 GitHub stars.

Is agent-beacon safe to use?

Yes. agent-beacon passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.

How do I install agent-beacon?

Clone the repository with "git clone https://github.com/Asymptote-Labs/agent-beacon" and add it to your Claude Code skills directory (see the Installation section above).

What programming language is agent-beacon written in?

agent-beacon is primarily written in Go. It is open-source under Asymptote-Labs on GitHub, so you can review or fork the full source.

Are there alternatives to agent-beacon?

Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh agent-beacon against similar tools.

Comments (0)

No comments yet. Be the first to share your thoughts!

ECC

by affaan-m

10

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

242,21936,702JavaScript
AI Agentsai-agentsanthropicclaude-code
View details
15

An agentic skills framework & software development methodology that works.

234,96620,863Shell
AI Agentsai-agentsbrainstorming
View details

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

185,94028,768JavaScript
AI Agentsai-agentsanthropicclaude-code
View details

cc-switch

by farion1231

3

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

128,8688,826Rust
AI Agentsclaude-codeai-tools
View details

claude-code

by anthropics

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

120,03119,897Shell
AI Agents
View details

Developers Also Liked

Based on votes and bookmarks from developers who liked this skill

ECC

by affaan-m

10

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

242,21936,702JavaScript
AI Agentsai-agentsanthropicclaude-code
View details
15

An agentic skills framework & software development methodology that works.

234,96620,863Shell
AI Agentsai-agentsbrainstorming
View details

n8n

by n8n-io

12

Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.

201,88160,308TypeScript
MCP Serversapisai-tools
View details

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

185,94028,768JavaScript
AI Agentsai-agentsanthropicclaude-code
View details

cc-switch

by farion1231

3

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

128,8688,826Rust
AI Agentsclaude-codeai-tools
View details