Skip to main content
Arize Skills are the primary interface for AI agents to work with Arize AX. Instead of agents calling APIs or CLIs blindly, skills give agents structured, domain-specific instructions so they can create datasets, run experiments, export traces, and optimize prompts correctly and safely. Works with Cursor, Claude Code, Codex, Windsurf, and 40+ other agents.

​
Quick Start

Make sure you have the AX CLI installed before getting started. 
npx skills add Arize-ai/arize-skills --skill ‘*’ --yes
This auto-detects your agent (Cursor, Claude Code, Codex, etc.) and symlinks skills into place.

​
Option 2: git clone

git clone https://github.com/Arize-ai/arize-skills.git
cd arize-skills
./install.sh --project ~/my-project
The installer detects installed agents and optionally installs the ax CLI. Use --global / -Global instead to install to ~/.<agent>/skills/.

​
Installer flags

FlagDescription
--project <dir>Target project directory for skill symlinks
--globalInstall to ~/.<agent>/skills/ instead (alternative to --project)
--copyCopy files instead of symlinking
--forceOverwrite existing skills
--skip-cliDon’t install ax CLI even if missing
--agent <name>Manually specify agent (cursor, claude, codex) — repeatable
--skill <name>Only install/uninstall specific skills — repeatable
--yesSkip confirmation prompts
--uninstallRemove previously installed skill symlinks
--listList all available skills and exit

​
Why Skills for Agents?

Agents (in Cursor, Claude Code, and similar environments) need to know when to use Arize and how to use it. Skills encode that knowledge:
  • When to invoke: Each skill has a clear description and trigger phrases (e.g., “creating or managing Arize datasets”, “downloading traces”, “running experiments”).
  • How to use AX: Skills document the ax CLI commands, concepts (datasets, experiments, traces, spans), and prerequisites (API key, space ID, project).

​
Available Arize Skills

The arize-skills repository provides skills that agents can load and use:
SkillDescription
arize-instrumentationSet up and configure tracing and instrumentation for applications (e.g., OpenTelemetry, Arize AX tracer registration, framework-specific instrumentors).
arize-linkGenerate deep links to traces, spans, and sessions in the Arize UI.
arize-datasetCreate, list, get, update, or delete datasets; append examples; export dataset data; create datasets from files.
arize-experimentCreate experiments, list runs, export experiment results, compare runs, and run evaluation workflows.
arize-traceExport traces and spans by trace ID, span ID, or session ID; debug LLM applications using trace data.
arize-prompt-optimizationOptimize or debug LLM prompts using production trace data, evaluations, and annotations; extract prompts from spans and run data-driven optimization.
Each skill defines concepts (e.g., trace vs. span vs. session), prerequisites (CLI, API key, space ID, project), and the exact ax commands and flags to use, so the agent can interface with Arize AX without guessing.

​
How Agents Use Skills

  1. Install or load skills in the agent environment using one of the Quick Start methods above.
  2. User asks in natural language (e.g., “Create a dataset for my eval”, “Export traces for session X”, “Run an experiment on this dataset”).
  3. Agent matches the request to a skill’s description and triggers.
  4. Agent follows the skill’s instructions: check prerequisites (ax --version, API key, space ID, project), then run the documented ax commands with the right flags and output handling.
Skills are the recommended way for agents to interface with Arize AX: they keep behavior consistent, reduce errors, and make it easy to add new capabilities (e.g., new skills or updated CLI usage) without changing agent code.

​
Prerequisites for Using Skills

Agents using Arize skills need the same setup as a human using the CLI:
  1. AX CLI installed: 
    # Preferred (isolated environment)
uv tool install arize-ax-cli
# or
pipx install arize-ax-cli
# Fallback
pip install arize-ax-cli
  2. API key — set up with a profile (ax profiles create) or via the ARIZE_API_KEY environment variable. Get your key from Arize app → API Keys.
  3. Space ID (from the Arize URL: /spaces/{SPACE_ID}/...) and project name as used in the Arize UI.
Skills tell the agent how to verify these (e.g., ax profiles show, ax projects list) and when to ask the user for missing values. See AX CLI overview for full CLI setup.

​
Updating Skills

  • npx path: npx skills update
  • git clone path: cd arize-skills && git pull (symlinks update automatically)

​
Related Resources

For questions or feedback on Arize skills, join the Arize community Slack.