agent-debugging
Debug and troubleshoot ElevenLabs conversational AI agents and Twilio calls. Use when diagnosing agent issues, analyzing failed calls, troubleshooting audio problems, investigating conversation breakdowns, reviewing error logs, or optimizing underperforming agents. Includes transcript analysis, error diagnosis, and performance troubleshooting.
npx skills add majiayu000/claude-skill-registry --skill agent-debugging --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Systematic debugging and troubleshooting framework for ElevenLabs Conversational AI + Twilio integration, offering diagnostic procedures, root cause analysis techniques, and resolution strategies to diagnose and fix agent issues, call quality, and context injection problems.
How it works
Follows a 5-step Debug Process: OBSERVE, REPRODUCE, ISOLATE, DIAGNOSE, FIX. Provides diagnostic steps and concrete commands/scripts for common issues:
- Agent Configuration: verify agent exists, voice validity, and agent config; inspect frontmatter and YAML; fix by creating/updating agent with validated config.
- Incorrect/Off-Topic Responses: retrieve full config, analyze system prompt quality, review conversation transcripts, and improve systemPrompt via an updateAgent call with a clarified, constraint-rich prompt.
- Audio & Call Quality: analyze transcripts for latency, pauses, and user clarifications; switch to faster models, adjust voice settings, and configure Twilio codec, jitter buffer, and echo cancellation; shorten system prompts to reduce processing time.
- Context & Integration: ensure lead context is correctly passed to conversations, verify context formatting, and update systemPrompt to reference leadData and propertyInfo; test by starting conversations with context and examining transcripts for lead name and address usage.
- Twilio Call Failures: check Twilio debugger, verify webhook accessibility, validate phone numbers, and review logs; fix common errors by ensuring valid E.164 numbers and proper TwiML responses.
When to use it
Invoke for debugging agent configuration issues, analyzing failed or low-quality calls, troubleshooting audio/latency problems, investigating conversation breakdowns, reviewing error logs, diagnosing context injection failures, identifying performance bottlenecks, and generating diagnostic reports.
What it can touch
Ops involve: starting/updating conversations and agents, reading agent and conversation data, and writing fixes via calls like:
- Use mcp__elevenlabs__elevenlabs_get_agent with: { "agentId": "your_agent_id" }
- Use mcp__elevenlabs__elevenlabs_get_voice with: { "voiceId": "voice_from_agent_config" }
- Read .claude/agents/elevenlabs-agent-manager.md
- Update agent via updateAgent(agentId, { ... })
- startConversation({ agentId, context })
- getConversation(conversationId)
- Grep logs/calling-service.log
- Twilio config edits including: router.post('/twiml/:conversationId', ...) and codec, jitterBufferSize, echoCancellation settings
Caveats
License: MIT. The skill outlines diagnostic steps and code samples but does not guarantee issue resolution or performance; outcomes depend on environment and proper data.
# Agent Debugging Skill Systematic debugging and troubleshooting framework for **ElevenLabs Conversational AI + Twilio** integration on the Next Level Real Estate platform. This skill provides diagnostic procedures, root cause analysis techniques, and resolution strategies. ## When to Use This Skill Invoke this skill when you need to: - ✅ Debug agent configuration issues - ✅ Analyze failed or poor-quality calls - ✅ Troubleshoot audio/latency problems - ✅ Investigate conversation breakdowns - ✅ Review and interpret error logs - ✅ Diagnose context injection failures - ✅ Identify performance bottlenecks - ✅ Generate diagnostic reports ## Debugging Methodology ### The 5-Step Debug Process ``` 1. OBSERVE → Gather symptoms and error data 2. REPRODUCE → Confirm issue is consistent 3. ISOLATE → Narrow down to specific component 4. DIAGNOSE → Identify root cause 5. FIX → Implement and verify solution ``` ## Common Issues & Solutions ### Issue Category 1: Agent Configuration #### Problem: Agent Not Starting/Responding **Symptoms:** - Agent status shows "inactive" or "error" - Conversation fails to start - No greeting heard on call - Error: "Agent not found" or "Invalid agent ID" **Diagnostic
- When to Use This Skill
- Debugging Methodology
- The 5-Step Debug Process
- Common Issues & Solutions
- Issue Category 1: Agent Configuration
- Issue Category 2: Audio & Call Quality
- Issue Category 3: Context & Integration
- Issue Category 4: Performance & Reliability
- Logging & Monitoring
- Comprehensive Logging Setup
- Monitoring Dashboard Queries
- Diagnostic Report Template
- Troubleshooting Checklist
- Resources
Use mcp__elevenlabs__elevenlabs_get_agent with: Expected output: Use mcp__elevenlabs__elevenlabs_get_voice with: Read .claude/agents/elevenlabs-agent-manager.md Look for: Use mcp__elevenlabs__elevenlabs_get_agent to retrieve full config Use mcp__elevenlabs__elevenlabs_get_conversation with conversationId Calculate time between user speech end and agent response start Grep logs/calling-service.log for: Use mcp__elevenlabs__elevenlabs_get_conversation
What does the agent-debugging skill do?
Debug and troubleshoot ElevenLabs conversational AI agents and Twilio calls. Use when diagnosing agent issues, analyzing failed calls, troubleshooting audio problems, investigating conversation breakdowns, reviewing error logs, or optimizing underperforming agents. Includes transcript analysis, error diagnosis, and performance troubleshooting.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill agent-debugging --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.