👁️Agent Observability Skills
Browse the best Agent Observability skills for Claude Code, Codex, Cursor and other AI agents.
260 skills · Page 3 of 11
langfuse-observability
★ 5LLM observability with self-hosted Langfuse 3.x - tracing, evaluation, monitoring, prompt management, and cost tracking
cc-conversation-analyzer
★ 5Comprehensive Claude Code conversation analysis skill for deep-diving into CC session logs.Use when analyzing exported Claude Code conversations to understand: project patterns, error rates,command failures, security risks, session duration, tool usage, and workflow efficiency.Triggers: "analyze conversation", "CC analysis", "conversation analysis", "session review","Claude Code logs", "analyze my sessions", "review CC usage", "conversation insights","what went wrong in my session", "session forensics", "CC forensics"
orchestration-logger
★ 5Unified logging system for AI orchestration workflows with current state metadata and execution logs in a single file
convex-agents-usage-tracking
★ 5Tracks LLM token consumption and usage metrics for billing, monitoring, and optimization. Use this to log token usage, calculate costs, generate invoices, and understand which agents or users consume the most resources.
convex-agents-playground
★ 5Sets up a web UI for testing, debugging, and developing agents without code. Use this to manually test agents, browse conversation history, and verify behavior in real-time.
convex-agents-debugging
★ 5Troubleshoots agent behavior, logs LLM interactions, and inspects database state. Use this when responses are unexpected, to understand context the LLM receives, or to diagnose data issues.
skill-performance-profiler
★ 5Analyzes skill usage patterns across conversations to track token consumption, identify heavy vs. lightweight skills, measure invocation frequency, detect co-occurrence patterns, and suggest consolidation opportunities. Use when the user asks to analyze skill performance, optimize skill usage, identify token-heavy skills, find consolidation opportunities, or review skill metrics.
human-in-the-loop
★ 5Instrument human approval workflows, feedback loops, and escalations
self-repair
★ 4Missing state triggers repair, not failure.
metrics
★ 4Collect and report metrics on agent usage from git history implementing the 8 key metrics defined in agent-metrics.md for measuring agent system health effectiveness and adoption.
install-skill-tracker
★ 4This skill installs a complete Claude Code skill tracking system using hooks to automatically log skill usage, execution duration, token usage, and user prompts for later analysis. Use this skill when setting up activity tracking in a Claude Code project to identify patterns, monitor costs, and discover opportunities for new skills.
ruby-integration
★ 4This skill is for writing integrations to the Ruby SDK. Claude acts as the engineer implementing LLM provider or agentic framework integrations. Use when adding support for OpenAI-like providers, Anthropic-like providers, or agent frameworks. Covers TDD workflow, comprehensive testing (streaming/non-streaming/tokens/multimodal), defensive coding, MCP validation, and StandardRB compliance.
session-chronicle
★ 4Excavate session logs for provenance tracking. TRIGGERS - who created, document finding, trace origin, session archaeology, provenance, ADR reference.
telemetry-pr-insights
★ 3Concise, PR-friendly telemetry reports scoped to current branch lifetime. For CI/automation.
telemetry-insights
★ 3Analyze AI coding session telemetry for usage patterns, token efficiency, and workflow optimization. Supports current session or historical time periods.
hook-event-architecture
★ 3Design hook-based event systems for ADW observability. Use when implementing real-time event broadcasting, creating hook pipelines, or building agent activity monitoring.
swimlane-visualization
★ 3Design swimlane UI patterns for visualizing ADW execution. Use when building observability dashboards, monitoring agent workflows, or creating real-time status displays.
standard-out-setup
★ 3Add console output and logging to make errors visible to agents. Standard out is a critical leverage point - without it, agents cannot see errors or understand application state. Use when agents fail silently, when debugging agentic workflows, or when setting up a new codebase for agentic coding.
multi-agent-observability
★ 3Build observability interfaces for multi-agent systems. Use when monitoring multi-agent execution, tracking agent metrics, implementing logging for parallel agents, or debugging agent workflows.
episode-log-steps
★ 3Log execution steps during episode execution for detailed tracking and learning. Use when recording tool usage, decision points, errors, or milestones during task execution.
sentry-setup-ai-monitoring
★ 2Setup Sentry AI Agent Monitoring in any project. Use this when asked to add AI monitoring, track LLM calls, monitor AI agents, or instrument OpenAI/Anthropic/Vercel AI/LangChain/Google GenAI. Automatically detects installed AI SDKs and configures the appropriate Sentry integration.
claude-mobile-cost-tracking
★ 2Use when implementing Claude API cost tracking, monitoring token usage, displaying cost metrics in Settings, or user asks about costs - calculates exact costs using $0.003/1k input and $0.015/1k output pricing with per-session aggregation
otel-observability
★ 2Implement OpenTelemetry-based observability for traces, metrics, and logs. Use for distributed tracing, Prometheus metrics, structured logging, and agent monitoring. Triggers on "OpenTelemetry", "OTEL", "tracing", "metrics", "observability", "Prometheus", "Grafana", "distributed tracing", or when implementing spec/007-observability.md.
cost-latency-optimizer
★ 2Reduces LLM costs and improves response times through caching, model selection, batching, and prompt optimization. Provides cost breakdowns, latency hotspots, and configuration recommendations. Use for "cost reduction", "performance optimization", "latency improvement", or "efficiency".