🔍 AI observability skill for Claude Code. Debug LangChain/LangGraph agents by fetching execution traces from LangSmith Studio directly in your terminal.

Claude Code session log → OpenTelemetry trace converter. Visualize AI agent behavior in Jaeger.

Local telemetry engine for AI coding tools and agents. Captures IDE-level traces calculates metrics from Claude Code and Cursor.

Automated harness evolution for AI agents. A Claude Code plugin that iteratively optimizes system prompts, routing, retrieval, and orchestration code using full-trace counterfactual diagnosis. Based on Meta-Harness (Lee et al., 2026).

A CLI tool for enabling agent traces for coding tools like Cursor, Claude Code etc.

AI-powered Chinese thesis polishing agent with 14 Claude Code skills: 10-dim AI trace detection, structure optimization, content expansion & quality assurance for LaTeX theses.中文论文智能修改助手 — 基于 Claude Code 的 14 个专业Skills，提供 10维 AI痕迹检测、结构优化、内容扩写与终稿质检的全流程论文打磨工具。

Governance framework for AI coding agents — enforces task traceability, structural gates, session continuity, and audit trails for Claude Code, Cursor, and Copilot.

Agent Trace poc for Claude Code

A comprehensive system that reproduces the Kiro editor workflow in Claude Code with EARS Hybrid TAD (Traceable Agentic Development) framework, delivering unambiguous requirements, behavioral contracts, and direct test translation.

Local-first Claude Code usage dashboard — sessions, agents, token usage, costs, traces

Hook kit for Claude Code and Gemini CLI that persists agent traces and working context into Beads for deterministic rehydration.

A general claude code loop template for having claude code write SDK agent where it recursively reviews its prompt and general scaffolding and modify itself and improve itself against real-valued function and/or past traces

Open-source toolkit that connects Problem-Driven AI methodology with GitHub Spec Kit. Think first, build second: two AI verification agents ensure everything the AI builds traces back to a real, validated problem. Works with Claude Code, GitHub Copilot, and Cursor.

Hierarchical multi-agent orchestration for Claude Code. Plans, delegates to parallel workers, reviews with Codex, and traces user flows across code boundaries. Think of it as a software engineering department in your terminal: a director plans the work, managers coordinate, and engineers (headless Claude Code sessions) execute.

Autonomous three-phase QA agent skill for Claude Code — analyzes, traces, and fixes code quality issues

Sentry AI Agent Monitoring plugin for Claude Code — traces sessions and tool calls as OpenTelemetry spans

🔍 Fetch execution traces from LangSmith Studio to debug LangChain and LangGraph agents effectively with Claude Code.

A structured workflow system for Claude Code that coordinates AI agents on complex tasks with traceable context.

Claude Code skill for analyzing Langfuse traces to diagnose Agent failures, identify error patterns, and generate actionable optimization recommendations.

🔍 Debug LangChain and LangGraph agents easily by fetching execution traces directly from LangSmith Studio in your terminal with Claude Code.

Claude Code Observability: route OTEL telemetry to LangFuse for session cost tracking, token analysis, tool profiling, and multi-agent trace correlation

ZAIA — AI Agent Team for Business-System Analysts. Claude Code skill for structured business analysis with 11 specialist agent roles, 6 phases, quality gates, and traceable artifact production.

OpenTelemetry instrumentation for the Claude Agent SDK. Auto-generates GenAI semantic convention spans, metrics, and tool traces via SDK hooks — zero code changes required.

Reduce Claude Code token usage 54–67%. MCP server for web engineers — AI agents find symbols, trace callers, and map hierarchies by name instead of reading files. Built for TypeScript. Works with Claude Code, Cursor, and Codex.

A Claude Code plugin providing eval and tracing observability for multi-agent skill pipelines. Uses per-agent JSONL ledgers and a shared orchestrator inbox for span logging, tool attestation, and trace lifecycle management. Includes a bootstrap manifest generator and LangSmith export for visualization.

Recursive AI self-improvement for ticket classification — an outer-loop Claude Opus 4.6 agent reads its own failure traces and writes improved harness code. Implementation of Meta-Harness (Lee et al., 2026). Built for FHNW course 1-26FS.W-B-BSC-KIdKIvDe.EN.