Cette page est affichée en anglais. Une traduction française est en cours.
SkillsMay 13, 2026·2 min de lecture

AgentSight — Zero-Instrumentation Agent Observability

AgentSight is an MIT eBPF tool that records LLM/agent traffic without SDK changes, helping you observe Claude Code or Gemini CLI interactions locally.

Script Depot
Script Depot · Community
Prêt pour agents

Installation avec revue préalable

Cet actif nécessite une revue. Le prompt copié demande un dry-run, affiche les écritures, puis continue seulement après confirmation.

Needs Confirmation · 64/100Policy : confirmer
Surface agent
Tout agent MCP/CLI
Type
Skill
Installation
Single
Confiance
Confiance : Established
Point d'entrée
Asset
Commande avec revue préalable
npx -y tokrepo@latest install 6fb9f9b6-7136-5255-b60f-b33eede8f3bb --target codex

Dry-run d'abord, confirmez les écritures, puis lancez cette commande.

Introduction

AgentSight is an MIT eBPF tool that records LLM/agent traffic without SDK changes, helping you observe Claude Code or Gemini CLI interactions locally.

Best for: developers who want local, low-overhead observability for agent tools without changing app code

Works with: Linux hosts with eBPF support, sudo privileges, Claude Code / Node / Python agent processes

Setup time: 10-15 minutes

Key facts (verified)

  • GitHub: 318 stars · 49 forks · pushed 2026-04-22.
  • License: MIT · owner avatar + repo URL verified via GitHub API.
  • README-verified entrypoint: docker run --privileged --pid=host --network=host \.

Main

  • Use it when instrumentation is too expensive: eBPF captures behavior at the system boundary without modifying your agent code.

  • Start by recording one process name (claude/node/python) and validate you can correlate events with failures or regressions.

  • Treat the trace as evidence: use it to debug latency spikes, unexpected network calls, or policy violations.

Source-backed notes

  • README states “Zero Instrumentation Required” and focuses on observing at the system boundary using eBPF.
  • README includes command examples to record Claude Code and gemini-cli processes via agentsight record -c ....
  • README suggests a local web UI at http://127.0.0.1:7395 for viewing recordings.

FAQ

  • Does it require code changes?: No—README emphasizes zero instrumentation.
  • Where does it run?: README examples target local processes on a host with sudo permissions.
  • Is it cross-platform?: It’s eBPF-based, so Linux support is the primary path (check README).
🙏

Source et remerciements

Source: https://github.com/eunomia-bpf/agentsight > License: MIT > GitHub stars: 318 · forks: 49

Fil de discussion

Connectez-vous pour rejoindre la discussion.
Aucun commentaire pour l'instant. Soyez le premier à partager votre avis.

Actifs similaires

Pipelock — MCP Firewall for Agent Egress

Run Pipelock as an agent firewall/proxy to scan MCP traffic for injection, secrets, SSRF, and risky tool chains; integrate with Claude Code fast.

SkillsCLI Tools
Script Depot

Hector — Self-Hosted Agent Runtime Binary

Run Hector as a self-hosted agent runtime: a single Go binary that serves an agent API and studio UI, keeping execution on your own infrastructure.

SkillsCLI Tools
Agent Toolkit

Codeg — Local AI Coding Workspace (Tauri/Web)

Run Codeg as a standalone server or desktop app to manage local agent workflows, MCP installs, and engineering loops; one-line install is available.

SkillsCLI Tools
Script Depot

Claude Swarm — Multi-Agent Orchestration with SDK

Python-based multi-agent orchestration built on Claude Agent SDK. Opus decomposes tasks, Haiku workers execute in parallel waves with real-time TUI dashboard and budget control.

SkillsCLI Tools
Agent Toolkit
◈Accueil🔍Rechercher👤Moi