Track and record real-time agent activity across MCP-enabled environments.
npx skills add https://github.com/belljun3395/agent-tracer --skill monitor-belljun3395
Please help me install this Agent Skill. Skill: monitor Source: https://github.com/belljun3395/agent-tracer/tree/main/.agents/skills/monitor Command: npx skills add https://github.com/belljun3395/agent-tracer --skill monitor-belljun3395
Agent Tracer 모니터링은 MCP 지원 환경에서 에이전트 활동을 실시간으로 관찰하고 로그를 체계적으로 기록하는 데 초점을 맞춥니다. 훅/플러그인이 없는 환경에서는 전체 이벤트를 수동으로 기록합니다.
Start the monitor server, then initiate a new monitoring task to begin capturing events during an agent session.
High-intent search queries and answers about installing and using this skill.
Monitor MCP agent activity by deploying a monitor-server and invoking the defined event API set to track real-time workflow actions. This captures events like task execution, tool usage, and terminal commands across MCP-enabled environments.
Yes, agent observability logging works with Cursor and Windsurf by manually recording events. While Claude Code and OpenCode support automatic logging via hooks, environments lacking plugins require manual event logging.
Yes, you can track agent thoughts and plans during an MCP session using monitor_thought and monitor_plan events. The defined event API set captures granular workflow stages including actions, verifications, and rules.
Start capturing agent events by launching the monitor-server and initiating a new monitoring task. This begins recording session activities, applying automatic hooks or manual logging based on your environment.
Automatic event logging uses hooks and plugins to collect agent activity in Claude Code and OpenCode, while manual logging requires recording each event by hand in environments like Cursor and Windsurf where automation is unavailable.
Without hooks or plugins, monitoring agent activity requires manual logging of the full event set, increasing overhead and risking incomplete traces. Environments lacking native automation cannot capture events automatically.
TeoSlayerDeploy a four-agent brand protection system with scanner, classifier, enforcer, and dashboard roles.
EdgeAppResolve agent sessions into structured JSON evidence manifests with transcripts and logs.
akillnessAutomate AI agent development governance with slash-command workflows across multiple coding platforms.
m2ai-portfolioDefine a canonical 10-event observability schema with shared agent_run_id for agent runs.
