A single, unified skill for Laminar: instrument code with tracing, run the Laminar CLI, query trace data with SQL, debug agents with record/replay, and migrate from other observability tools.

npx skills add lmnr-ai/laminar-skills --skill laminar

• "Set up Laminar in this repo."
• "Instrument this codebase with Laminar and choose the most useful spans."
• "Show me a Laminar example that produces a trace in under 2 minutes."
• "Use the Laminar CLI to query the slowest spans in the last 24 hours."
• "Migrate our Langfuse tracing to Laminar."
• "Debug my agent with the Laminar debugger: record a run, find the bad LLM call, and replay up to it."

The skill routes from a single SKILL.md to focused, one-level-deep reference files:

• skills/laminar/SKILL.md — mental model, workflow, and task router
• skills/laminar/references/instrumentation-typescript.md — TypeScript/JS/Next.js setup, spans, context, privacy
• skills/laminar/references/instrumentation-python.md — Python setup, spans, context, privacy
• skills/laminar/references/quickstart.md — minimal demo trace (Node + Python) and troubleshooting
• skills/laminar/references/cli.md — lmnr-cli and the SDK-bundled lmnr CLI (sql, datasets, dev, eval)
• skills/laminar/references/sql-query-api.md — SQL Query API and ClickHouse query patterns
• skills/laminar/references/migration.md — migrate from Langfuse, LangSmith, Helicone, or OpenTelemetry
• skills/laminar/references/debugging.md — agent debugger: record under LMNR_DEBUG, inspect with SQL, replay cached LLM calls, annotate sessions

• You will need a Laminar project API key (dashboard → Settings → Project API Keys) to send traces or execute queries.