livekit-agents
Build voice AI agents with LiveKit Cloud and the Agents SDK. Use when the user asks to "build a voice agent", "create a LiveKit agent", "add voice AI", "implement handoffs", "structure agent workflows", or is working with LiveKit Agents SDK. Provides opinionated guidance for the recommended path: LiveKit Cloud + LiveKit Inference. REQUIRES writing tests for all implementations.
60
70%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Advisory
Suggest reviewing before use
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/livekit-agents/SKILL.mdQuality
Discovery
89%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong skill description with excellent trigger term coverage and completeness. The explicit 'Use when...' clause with multiple natural phrases makes it highly discoverable. The main weakness is that the capability description could be more specific about concrete actions beyond 'build' and 'implement'.
Suggestions
Add 2-3 more specific concrete actions to improve specificity, e.g., 'configure speech-to-text pipelines, manage conversation state, handle voice interruptions'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (voice AI agents, LiveKit) and mentions some actions like 'build', 'implement handoffs', 'structure agent workflows', but doesn't list multiple concrete specific actions like 'create voice pipelines, configure STT/TTS, handle interruptions'. | 2 / 3 |
Completeness | Clearly answers both what ('Build voice AI agents with LiveKit Cloud and the Agents SDK') and when (explicit 'Use when...' clause with multiple trigger scenarios). Also includes additional context about recommended path and testing requirements. | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural trigger terms users would say: 'build a voice agent', 'create a LiveKit agent', 'add voice AI', 'implement handoffs', 'structure agent workflows', 'LiveKit Agents SDK' - these are phrases users would naturally use. | 3 / 3 |
Distinctiveness Conflict Risk | Very distinct niche with LiveKit-specific terminology ('LiveKit Cloud', 'LiveKit Inference', 'Agents SDK') that clearly differentiates it from generic voice or AI skills. Unlikely to conflict with other skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
50%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill provides solid architectural guidance and design principles for LiveKit voice agents, with appropriate emphasis on documentation verification and testing requirements. However, it suffers from repetition (the 'never trust model memory' message appears excessively), lacks executable code examples, and provides more meta-guidance about consulting documentation than concrete implementation patterns. The skill tells developers how to think but doesn't give them enough to actually do.
Suggestions
Add at least one complete, executable test example showing the actual testing framework syntax and assertions rather than just describing what to test
Consolidate the 'never trust model memory' warnings into a single prominent section rather than repeating throughout the document
Include a concrete agent implementation example (even minimal) showing the actual SDK patterns for creating an agent with LiveKit Inference
Add explicit validation checkpoints to the development workflow (e.g., 'Run agent locally and verify greeting works before adding tools')
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill contains some unnecessary repetition (the 'never trust model memory' rule is stated 4+ times) and verbose explanations of concepts Claude likely knows (what latency is, why tests matter). However, it avoids explaining basic concepts like what voice AI is and stays focused on LiveKit-specific guidance. | 2 / 3 |
Actionability | The skill provides some concrete guidance (environment variable setup, directory structure) but lacks executable code examples. Most instructions are meta-guidance ('consult documentation', 'use MCP') rather than copy-paste ready implementations. The testing section mentions patterns but provides no actual test code. | 2 / 3 |
Workflow Clarity | The mandatory checklist provides clear sequencing, and the test-driven development process has steps. However, the actual agent development workflow lacks explicit validation checkpoints—there's no 'verify your agent works before proceeding' step, and the handoffs/tasks section describes concepts without concrete implementation sequences. | 2 / 3 |
Progressive Disclosure | The skill is well-organized with clear sections and headers, but it's somewhat monolithic at ~300 lines. It references external documentation appropriately but doesn't split its own content into separate files (e.g., testing guidance could be TESTING.md, architecture principles could be ARCHITECTURE.md). The structure is good but could be better layered. | 2 / 3 |
Total | 8 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- livekit/agent-skills
- Commit
d86b645
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.