Skip to main content
hc
Submit
skillsSource-backedReview first Safety ✓ Privacy ✓

LiveKit Agent Skills

Official LiveKit Agent Skills for AI coding agents building low-latency voice AI, LiveKit Agents workflows, handoffs, mandatory tests, and simulation scenario suites.

by LiveKit·added 2026-06-18·
Claude CodeCodexWindsurfGeminiCursorCLI
HarnessClaude CodeCodexWindsurfGeminiCursorCLI
Level:expertType:capability-packVerified:validated
Review first — review before installing

Open the source and read safety notes before installing.

Safety notes

  • The livekit-agents skill intentionally pushes agents toward implementation work for voice AI systems that can join realtime rooms, call tools, speak to users, and route calls; generated code still needs human review.
  • The skill requires tests for agent behavior, but tests do not prove latency, safety, consent, telephony legality, privacy, or production readiness by themselves.
  • The livekit-simulations skill includes private-beta caveats for simulation commands and requires current CLI help or docs verification before running `lk agent simulate`.
  • Do not let a coding agent invent LiveKit API signatures from memory; the skill repeatedly requires MCP/docs verification because the SDK changes quickly.
  • Voice agent handoffs, tasks, tool calls, and simulation scenarios can influence real user conversations if deployed; validate in staging rooms before production.

Privacy notes

  • LiveKit voice agent work can involve audio, video, transcripts, room metadata, participant identities, phone call details, test personas, tool inputs, tool outputs, and logs.
  • The skills are prompt/instruction assets, but the implementations they guide may send data to LiveKit, STT providers, LLM providers, TTS providers, MCP servers, telephony providers, and observability backends.
  • Keep LIVEKIT_API_SECRET, provider keys, SIP credentials, room tokens, recordings, transcripts, and generated scenario files containing sensitive business logic out of prompts, public issues, screenshots, and committed configs.
  • The simulations skill says scenario generation reads the user's local agent code and should not upload that code; preserve that local-only boundary when using it.

Prerequisites

  • AI coding agent or skill installer compatible with Agent Skill-style repositories.
  • LiveKit Agents project or planned voice AI implementation.
  • LiveKit Cloud project credentials when using the recommended LiveKit Cloud path.
  • LiveKit Docs MCP server or current docs access for API signatures, CLI commands, model support, deployment, and configuration facts.
  • Testing environment for LiveKit Agents behavior, including local unit tests or simulation runs where supported.

Schema details

Install type
package
Reading time
6 min
Difficulty score
72
Troubleshooting
Yes
Breaking changes
No
Source repository stats
Scope
Source repo
Package metadata
Skill and platform metadata
Skill type
capability-pack
Skill level
expert
Verification
validated
Verified at
2026-06-18
Retrieval sources
https://github.com/livekit/agent-skillshttps://raw.githubusercontent.com/livekit/agent-skills/main/README.mdhttps://raw.githubusercontent.com/livekit/agent-skills/main/skills/livekit-agents/SKILL.mdhttps://raw.githubusercontent.com/livekit/agent-skills/main/skills/livekit-simulations/SKILL.mdhttps://raw.githubusercontent.com/livekit/agent-skills/main/LICENSE
Tested platforms
ClaudeCodexCursorGeneric AGENTS
PlatformSupportInstall path
claude-codeNative.claude/skills/<skill-name>/SKILL.md
codexNative.agents/skills/<skill-name>/SKILL.md
windsurfNative.windsurf/skills/<skill-name>/SKILL.md
geminiNative.gemini/skills/<skill-name>/SKILL.md or .agents/skills/<skill-name>/SKILL.md
cursorAdapter.cursor/rules/<skill-name>.mdc
cliManualAGENTS.md or tool-specific context file
Tool listing metadata
Full copyable content
npx skills add livekit/agent-skills

# Install one focused skill when your installer supports --skill
npx skills add livekit/agent-skills --skill livekit-agents

About this resource

LiveKit Agent Skills

livekit/agent-skills is LiveKit's official skill repository for AI coding agents that build and test realtime voice AI systems. It packages two reusable skills:

  • livekit-agents: opinionated guidance for building low-latency LiveKit Agents implementations with workflow design, handoffs, tasks, and mandatory testing.
  • livekit-simulations: guidance for generating local simulation scenarios from an agent's own code and user-stated test focus, then running those scenarios through LiveKit's simulation tooling where available.

Use this when the task is about voice AI architecture, LiveKit Agents project structure, handoffs, test strategy, or simulation coverage. Use the separate LiveKit Agents framework entry when evaluating the runtime/package itself.

Knowledge Freshness

The skill repository intentionally avoids freezing LiveKit API signatures into the prompt. LiveKit Agents, LiveKit Cloud, LiveKit Inference, CLI commands, simulation support, model providers, and deployment guidance can change quickly.

Prefer current docs.livekit.io, the LiveKit Docs MCP server, the installed SDK version, and lk ... --help output over model memory. The skills provide durable behavioral guidance; live docs provide exact API facts.

Retrieval Sources

Use these sources to ground an agent session:

  • Upstream livekit/agent-skills README for install, available skills, design philosophy, and MCP pairing.
  • skills/livekit-agents/SKILL.md for voice AI architecture, handoffs, tasks, testing requirements, and LiveKit Cloud assumptions.
  • skills/livekit-simulations/SKILL.md for local scenario generation, simulation beta caveats, risk checklists, and coverage enforcement.
  • Upstream license and repository metadata for source trust and distribution.
  • LiveKit Docs MCP server or current docs.livekit.io pages for API signatures and command behavior.

Core Workflow

  1. Install the skill repository with the supported skill installer.
  2. Confirm whether the task is building a LiveKit agent, modifying an existing agent, or testing an existing agent.
  3. Load livekit-agents for architecture and implementation guidance, or livekit-simulations when the user wants scenarios, stress tests, or regression coverage.
  4. Verify all LiveKit SDK, CLI, MCP, deployment, and model-provider details against current docs before writing code.
  5. Include tests or simulation scenarios before calling the implementation done.

Capability Scope

This capability pack helps AI coding agents approach LiveKit voice AI work with the right operating model: latency-first architecture, small contexts, focused tools, handoffs, tasks, current-docs verification, and test coverage.

It does not provide LiveKit credentials, create a LiveKit Cloud project, replace the LiveKit Docs MCP server, guarantee current SDK signatures, certify a voice agent for production, or resolve consent, telecom, privacy, and recording requirements.

Production Rules

  • Do not ship a LiveKit voice agent without behavior tests or reviewed simulation coverage.
  • Do not rely on model memory for LiveKit APIs; verify signatures, CLI flags, and deployment steps through current docs or MCP.
  • Keep voice agent context, tools, and handoff scopes small enough for realtime latency.
  • Keep LiveKit secrets, provider keys, room tokens, SIP credentials, recordings, transcripts, and generated scenarios out of public artifacts.
  • Treat simulation support as version- and account-dependent when the upstream skill marks it private beta.

Install

Install the full skill repository:

npx skills add livekit/agent-skills

When your installer supports installing a focused skill, install only the architecture skill for implementation tasks:

npx skills add livekit/agent-skills --skill livekit-agents

Skill Inventory

Skill Use It For
livekit-agents Building or modifying LiveKit Agents implementations, especially voice AI workflows, handoffs, task structure, latency-sensitive design, and mandatory behavior tests
livekit-simulations Generating local test scenarios, risk checklists, scenario YAML, and simulation runs for existing LiveKit voice or chat agents

Pair With LiveKit Docs MCP

The upstream README explicitly positions these skills alongside the LiveKit Docs MCP server:

  • Skills teach the agent how to approach voice AI work.
  • MCP/docs provide current facts: API signatures, configuration options, command flags, deployment steps, examples, and model-provider details.

That division matters. A model can remember old LiveKit APIs and still produce plausible code. The skill tells the agent to verify rather than guess.

Source Review

Verified on 2026-06-18:

  • The upstream README describes livekit/agent-skills as reusable skills for AI coding agents building with LiveKit.
  • The README lists livekit-agents and livekit-simulations as the available skills and documents installation with npx skills add livekit/agent-skills.
  • The README states the skills are designed to work with the LiveKit Docs MCP server, with skills providing behavioral guidance and MCP providing factual API information.
  • skills/livekit-agents/SKILL.md describes LiveKit Cloud-oriented voice agent guidance, current-docs verification, latency-first architecture, handoffs, tasks, and mandatory tests.
  • skills/livekit-simulations/SKILL.md describes local scenario generation, risk checklist coverage, authored scenarios, lk agent simulate, and private beta caveats for simulation command availability.
  • GitHub reports MIT licensing and current activity for livekit/agent-skills.

Duplicate Check

Checked current content/skills/, content/tools/, content/mcp/, content/agents/, guides, open and closed pull requests, and repository-wide content for livekit/agent-skills, LiveKit Agent Skill, livekit-agents skill, livekit-simulations, LiveKit Docs MCP, voice AI agent skills, and npx skills add livekit/agent-skills. The existing LiveKit Agents tools entry mentions the skill as recommended companion context, but no dedicated LiveKit Agent Skills entry, source URL duplicate, or open duplicate PR was found.

Disclosure

Editorial listing. No paid placement or affiliate link is used. The skill repository is MIT licensed and maintained by LiveKit.

#livekit#skills#voice-ai#ai-agents#testing#simulations#mcp#claude-code

Source citations

Add this badge to your README

Show that LiveKit Agent Skills is listed on HeyClaude. Paste this Markdown into your README — it renders the badge and links back to this page.

Listed on HeyClaude
[![Listed on HeyClaude](https://heyclau.de/badge/skills/livekit-agent-skills.svg)](https://heyclau.de/entry/skills/livekit-agent-skills)

How it compares

LiveKit Agent Skills side by side with 3 alternatives on trust, install, platform support, and disclosed safety notes — all from reviewed registry metadata.

FieldLiveKit Agent Skills

Official LiveKit Agent Skills for AI coding agents building low-latency voice AI, LiveKit Agents workflows, handoffs, mandatory tests, and simulation scenario suites.

Open dossier 		Supabase Agent Skills

Official Supabase Agent Skills for AI coding agents working with Supabase Database, Auth, Edge Functions, Realtime, Storage, Vectors, CLI, MCP, RLS, migrations, and Postgres performance.

Open dossier 		.NET Agent Skills

Microsoft .NET team skill marketplace for AI coding agents working on .NET, C#, ASP.NET Core, Blazor, MAUI, diagnostics, MSBuild, NuGet, upgrades, tests, AI workflows, RAG pipelines, and C# MCP servers.

Open dossier 		Dart Agent Skills

Official Dart team Agent Skills for AI coding agents working on Dart unit tests, CLI apps, coverage, runtime errors, mocks, package conflicts, static analysis, Native Assets, FFI, ffigen, and pattern matching.

Open dossier
Trust
Install riskReview firstReview firstReview firstReview first
Notes Safety ✓ Privacy ✓ Safety ✓ Privacy ✓ Safety ✓ Privacy ✓ Safety ✓ Privacy ✓
Categoryskillsskillsskillsskills
Sourcesource-backedsource-backedsource-backedsource-backed
AuthorLiveKitSupabase.NET Team at MicrosoftDart Team
Added2026-06-182026-06-182026-06-182026-06-18
Platforms
Claude CodeCodexWindsurfGeminiCursorCLI
Claude CodeCodexWindsurfGeminiCursorCLI
Claude CodeCodexWindsurfGeminiCursorCLIVS Code
Claude CodeCodexWindsurfGeminiCursorCLI
Source repo————
Safety notes✓The livekit-agents skill intentionally pushes agents toward implementation work for voice AI systems that can join realtime rooms, call tools, speak to users, and route calls; generated code still needs human review. The skill requires tests for agent behavior, but tests do not prove latency, safety, consent, telephony legality, privacy, or production readiness by themselves. The livekit-simulations skill includes private-beta caveats for simulation commands and requires current CLI help or docs verification before running `lk agent simulate`. Do not let a coding agent invent LiveKit API signatures from memory; the skill repeatedly requires MCP/docs verification because the SDK changes quickly. Voice agent handoffs, tasks, tool calls, and simulation scenarios can influence real user conversations if deployed; validate in staging rooms before production.✓The skills guide agents toward Supabase schema, auth, RLS, migration, MCP, and Postgres work that can affect real user data; generated SQL and policies need human review. The Supabase skill explicitly treats RLS, exposed schemas, service-role keys, JWT claims, views, storage policies, and SECURITY DEFINER functions as security-sensitive areas. Do not let an agent apply migrations or production SQL just because the skill suggests a workflow; run advisors, inspect generated SQL, and verify against the intended environment. The repository includes Supabase MCP configuration for docs-only MCP access. Project-scoped MCP or database access still needs proper authentication, least privilege, and review.✓.NET build, test, upgrade, package, template, publish, and migration tasks can modify project files, lock files, generated code, packages, app settings, and deployment artifacts. Diagnostics skills may suggest collecting traces, dumps, counters, crash data, MSBuild binlogs, or performance profiles; collect those artifacts only with explicit approval and storage controls. MCP server skills can expose local code, files, APIs, credentials, or production services as callable tools; review tool descriptions, parameter validation, authorization, and transport choice before connecting clients. NuGet and publish workflows can push packages or artifacts to public or private feeds; verify package IDs, versions, API keys, feed targets, and release policy before publishing. Upgrade and modernization guidance should be verified against each application's framework support window, deployment target, package compatibility, and rollback plan.✓Dart skills can modify source code, tests, pubspec dependencies, analysis options, generated mocks, coverage output, native hooks, FFI bindings, and package constraints. `dart fix --apply`, formatter runs, mock generation, build_runner, ffigen, and Native Assets hooks can create broad mechanical diffs; inspect changes before committing. Native Assets and FFI workflows may compile C/C++ code, run local toolchains, download precompiled binaries, and link dynamic or static libraries. Downloaded binaries should use cryptographic hash validation and offline fallbacks as the skill describes. Package conflict fixes can upgrade or downgrade dependencies and affect transitive package behavior across Dart and Flutter projects.
Privacy notes✓LiveKit voice agent work can involve audio, video, transcripts, room metadata, participant identities, phone call details, test personas, tool inputs, tool outputs, and logs. The skills are prompt/instruction assets, but the implementations they guide may send data to LiveKit, STT providers, LLM providers, TTS providers, MCP servers, telephony providers, and observability backends. Keep LIVEKIT_API_SECRET, provider keys, SIP credentials, room tokens, recordings, transcripts, and generated scenario files containing sensitive business logic out of prompts, public issues, screenshots, and committed configs. The simulations skill says scenario generation reads the user's local agent code and should not upload that code; preserve that local-only boundary when using it.✓Supabase tasks can involve schemas, SQL, RLS policies, migrations, auth settings, JWT claims, storage paths, Edge Function code, logs, and customer data. Keep SUPABASE_ACCESS_TOKEN, service_role keys, database passwords, JWT secrets, project refs, connection strings, OAuth secrets, and private schema dumps out of prompts, screenshots, public PRs, and committed configs. When using Supabase MCP or CLI tools, the connected agent may see project metadata, database structure, logs, or SQL results depending on granted permissions. Docs-only MCP access is safer than project MCP access, but user queries and docs snippets can still be forwarded into the configured model provider.✓.NET repositories may contain connection strings, appsettings secrets, user secrets, certificates, environment variables, telemetry keys, logs, traces, dumps, package credentials, and production data. MSBuild binlogs, crash dumps, profiler output, and test artifacts can contain source paths, dependency graphs, request data, exception payloads, configuration values, and environment details. MCP servers created with these skills may forward prompts and tool inputs to local processes, HTTP services, databases, cloud APIs, or third-party model providers depending on the implementation. Keep private NuGet credentials, signing keys, deployment secrets, customer data, dumps, and proprietary source out of public prompts, issues, pull requests, and shared artifacts.✓Dart projects may contain private package names, API schemas, service URLs, test fixtures, generated mocks, coverage paths, stack traces, native library names, and build logs. Runtime-error workflows can expose stack traces, local paths, active app state, logs, and user-input data from the running program. FFI and Native Assets workflows may include proprietary C/C++ source, platform-specific build settings, binary hashes, and downloaded artifact URLs. Keep credentials, private package registries, source paths when sensitive, proprietary native code, test data, and generated coverage reports out of public prompts, issues, PRs, and examples.
Prerequisites
  • AI coding agent or skill installer compatible with Agent Skill-style repositories.
  • LiveKit Agents project or planned voice AI implementation.
  • LiveKit Cloud project credentials when using the recommended LiveKit Cloud path.
  • LiveKit Docs MCP server or current docs access for API signatures, CLI commands, model support, deployment, and configuration facts.
  • AI coding agent or skill installer compatible with Agent Skills repositories.
  • Supabase project, local Supabase app, or planned Supabase integration.
  • Current Supabase docs access, Supabase CLI help, or Supabase MCP server access for exact commands and API details.
  • Database access appropriate to the task when validating SQL, RLS, migrations, advisors, or performance fixes.
  • An AI coding assistant or skill host that supports Agent Skills, plugin marketplaces, or direct skill installation.
  • .NET SDK and project-local build/test tooling appropriate for the repository being edited.
  • For MCP server work, the official C# MCP SDK, MCP project templates, and a target transport choice such as stdio or HTTP.
  • For diagnostics, permission to collect traces, dumps, logs, counters, binlogs, or test output from the target environment.
  • Dart SDK and project-local tooling installed for the package being edited.
  • An AI coding assistant or skill host compatible with Agent Skills, preferably using the standard `.agents/skills` install target.
  • Project `pubspec.yaml`, `analysis_options.yaml`, test structure, package constraints, and platform targets available before modifying code.
  • For FFI or Native Assets work, compatible host compiler toolchains, package hooks, native source, binding generation configuration, and platform targets.
Install
npx skills add livekit/agent-skills
npx skills add supabase/agent-skills
codex plugin marketplace add dotnet/skills
npx skills add dart-lang/skills --skill '*' --agent universal --yes
Config————
Citations
ClaimUnclaimedUnclaimedUnclaimedUnclaimed

Signals

Loading live community signals…

More like this, weekly

A short, calm digest of reviewed Claude resources. Unsubscribe any time.