Parseable Skills

Agent skills for adding Parseable observability to LLM applications and AI agents.

Skills

Skill	Description
parseable-instrument	Add, audit, and validate Parseable tracing/logging for LLM apps and AI agents using OpenTelemetry, framework integrations, and collector-based export.

Installation

npx (recommended)

npx skills add parseablehq/skills --skill "parseable-instrument" --global

Manual symlink

git clone https://github.com/parseablehq/skills.git ~/parseable-skills
ln -s ~/parseable-skills/skills/parseable-instrument ~/.claude/skills/parseable-instrument

Prerequisites

You need a Parseable instance and credentials:

export PARSEABLE_ENDPOINT=https://your-parseable-instance.example.com
export PARSEABLE_USERNAME=your-username
export PARSEABLE_PASSWORD=your-password
export PARSEABLE_STREAM=genai-traces

Use an API key or Basic auth according to your deployment's access model. Create the target dataset before exporting, or change PARSEABLE_STREAM to an existing dataset.

Usage

Once installed, the agent will automatically use the skill when relevant, for example:

• Adding Parseable tracing to an existing AI agent.
• Auditing LLM instrumentation for model names, token usage, errors, and span hierarchy.
• Setting up OpenTelemetry Collector export to Parseable.
• Choosing framework integrations for OpenAI, Anthropic, LiteLLM, OpenRouter, vLLM, LangChain, LlamaIndex, AutoGen, CrewAI, DSPy, n8n, or other OpenTelemetry-instrumented systems.
• Validating traces, logs, datasets, and dashboards in Parseable.

License

MIT