Agent Tracing
Agent Tracing is a zero-config CLI tool that captures agent execution snapshots during development and lets you inspect them for debugging. Query specific steps, view tool invocations and results, examine LLM messages and context, or pull remote production traces by operation ID. All inspection flags work on local and remote snapshots alike.
Agent Tracing provides CLI inspection of agent execution snapshots to debug step flows, tool calls, and LLM interactions.
AI-generated summary based on this skill's SKILL.md
Decision gist · record as of 2026-07-28
Agent Tracing provides CLI inspection of agent execution snapshots to debug step flows, tool calls, and LLM interactions. Agent Tracing is a zero-config CLI tool that captures agent execution snapshots during development and lets you inspect them for debugging. Query specific steps, view tool invocations and results, examine LLM messages and context, or pull remote production traces by operation ID. All inspection flags work on local and remote snapshots alike.
Use it when
- Agent Tracing captures execution snapshots automatically during development.
- Yes.
Install
lobehub/lobehub/agent-tracing · repository language: TypeScript
generated, unverified - the skill's exact subdirectory could not be determined; check the repository on GitHub
Open directory. Skills are indexed for reading, not audited. Review a skill's body before installing it.
Frequently asked questions
AI-generated answers based on this skill's SKILL.md and metadata
What is Agent Tracing and how do I use it?
Agent Tracing is a zero-config CLI tool that captures agent execution snapshots during development and lets you inspect them for debugging. It records tool invocations, results, LLM messages, and context so you can examine exactly what your agent did at each step without manual instrumentation.
How do I trace agent execution and debug agent behavior?
Agent Tracing captures execution snapshots automatically during development. Use the CLI to query specific steps, view tool invocations and their results, examine LLM messages and context, and inspect the complete execution flow. All inspection flags work on both local snapshots and remote production traces pulled by operation ID.
Can Agent Tracing monitor agent calls and interactions?
Yes. Agent Tracing records and displays all agent calls and interactions including tool invocations, LLM message exchanges, and context data. You can inspect these interactions locally during development or retrieve remote production traces by operation ID to monitor how your agent behaved in real scenarios.
Does Agent Tracing help visualize agent workflow and request paths?
Agent Tracing captures agent execution snapshots that let you inspect the complete workflow and request paths. You can query specific steps and view the sequence of tool invocations, LLM interactions, and results to understand exactly how requests flow through your agent.
How can I diagnose agent performance and runtime issues?
Agent Tracing provides diagnostic tracing capabilities to help identify performance bottlenecks and runtime issues. Inspect execution snapshots to see where time is spent, which tools are called, how LLM interactions occur, and what context is available at each step—all with zero-config setup.
Can I access remote production traces with Agent Tracing?
Yes. Agent Tracing lets you pull remote production traces by operation ID and inspect them using the same CLI tools you use for local snapshots. All inspection flags work on both local development snapshots and remote production traces.
Let your AI agent find skills like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 56,283 agent skills by what they can do, searchable in plain language.
wish › “Debug and trace agent execution flows and behavior”
Give your agent the search over MCP, or paste the wish link into any chat. No install? Search from any chat →
Related skills
Register custom handlers at 16 lifecycle points to observe and control agent behavior. Hooks fire before and after steps, tool calls, human approvals, context compression, and sub-agent invocations, with support for tool mocking and error handling. Handlers run sequentially per operation and auto-clean on completion.
Builtin Tool is a guide for authoring agent-callable tool packages in LobeHub. It covers the five architectural layers—manifest and types, execution runtime, frontend executor, client UI, and registry wiring—along with design principles, file organization, and a pre-flight checklist for new tools.
This skill provides a curated reference guide to LobeHub's open-source monorepo, detailing the directory structure, technology stack (Next.js 16, React 19, TypeScript, PostgreSQL), and key package locations. It maps UI components, SPA pages, state management, API routers, and database layers to help developers understand the codebase organization and find specific functionality quickly.
agentcore-investigation lets you troubleshoot Bedrock AgentCore runtime sessions by querying CloudWatch Logs Insights with structured and glob-style parsing. It resolves session-to-trace relationships through OpenTelemetry span correlation, filters instrumentation noise, and surfaces errors, tool invocations, token usage, and performance metrics in timeline format.
Agent Signal enables you to implement event-driven background work for agents without blocking foreground chat interactions. The system normalizes incoming events into sources, derives semantic signals from them, and executes concrete actions—all orchestrated through a consistent pipeline with built-in observability and durable self-iteration support.
Trace systematically investigates why something happened by generating competing hypotheses and evaluating evidence for and against each one. It's built for ambiguous causal questions—runtime bugs, performance regressions, and architecture issues—where you need to explain the most likely explanation. The skill structures your investigation through observation, hypothesis generation, evidence gathering, and discriminating probes.
More skills Agent Browser (NOASSERTION)