npx claudepluginhub arize-ai/arize-claude-code-plugin --plugin arize-platformThis skill uses the workspace's default tool permissions.
Manage projects in the Arize AI platform using the `ax` CLI.
Manages Arize AI datasets via ax CLI: list with pagination, get details, create from CSV/JSON/Parquet files, delete, export data, extract IDs with jq. For Arize ML platform ops.
Manages Arize ML platform resources like models, monitors, prompts, evaluators, dashboards, spaces via arize_toolkit CLI. Lists, creates, updates, deletes resources, configures profiles, handles admin tasks from terminal.
Creates, runs, and analyzes Arize experiments for evaluating and comparing AI model performance using ax CLI. Useful for experiment CRUD, result comparison, and model benchmarking.
Share bugs, ideas, or general feedback.
Manage projects in the Arize AI platform using the ax CLI.
The user must have:
pip install arize-ax-cli)ax config init)ax projects list
Options:
--space-id <id> - Space ID to list projects from (uses config default if not set)--limit, -n <count> - Maximum number of projects to return (default: 15)--cursor <token> - Pagination cursor for next page--output, -o <format> - Output format: table (default), json, csv, parquet, or a file path--profile, -p <name> - Configuration profile to use--verbose, -v - Enable verbose logsExamples:
# List projects (default table format)
ax projects list
# List as JSON
ax projects list --output json
# List from a specific space
ax projects list --space-id sp_abc123
# Limit results
ax projects list -n 5
# Use production profile
ax projects list --profile production
Extracting Project IDs:
# Get all project IDs and names as JSON
ax projects list --output json | jq '.[] | {id: .id, name: .name}'
# Find a project ID by name
ax projects list --output json | jq -r '.[] | select(.name == "My Project") | .id'
# Save project ID to a variable
PROJECT_ID=$(ax projects list --output json | jq -r '.[] | select(.name == "My Project") | .id')
echo "Found project: $PROJECT_ID"
Without jq (using grep):
# Find project by name
ax projects list --output json | grep -B 1 '"name": "My Project"' | grep "id" | cut -d'"' -f4
The CLI commands (get, delete) require a project ID, not a name. When a user refers to a project by name, resolve the ID first using projects list:
ax projects list --output jsonIf no exact match is found, check for partial or case-insensitive matches and confirm with the user before proceeding. If multiple matches exist, present the options and ask the user to choose.
# Example: user asks to "get the ML Experiments project"
PROJECT_ID=$(ax projects list --output json | jq -r '.[] | select(.name == "ML Experiments") | .id')
if [ -z "$PROJECT_ID" ]; then
echo "Project not found. Available projects:"
ax projects list --output json | jq '.[] | {id: .id, name: .name}'
else
ax projects get "$PROJECT_ID"
fi
Retrieve information about a specific project:
ax projects get <project-id>
Arguments:
id (required) - The project IDOptions:
--output, -o <format> - Output format: table (default), json, csv, parquet, or a file path--profile, -p <name> - Configuration profile to use--verbose, -v - Enable verbose logsExamples:
# Get project details
ax projects get proj_abc123
# Get as JSON
ax projects get proj_abc123 --output json
# Get from production environment
ax projects get proj_abc123 --profile production
Create a new project in a space:
ax projects create --name <name> --space-id <space-id>
Options:
--name, -n <name> (required) - Project name (prompted interactively if not provided)--space-id <id> (required) - Space ID to create the project in (prompted interactively if not provided)--output, -o <format> - Output format: table (default), json, csv, parquet, or a file path--profile, -p <name> - Configuration profile to use--verbose, -v - Enable verbose logsExamples:
# Create with all options specified
ax projects create --name "ML Experiments" --space-id sp_abc123
# Create interactively (prompts for name and space-id)
ax projects create
# Create and output as JSON
ax projects create --name "Staging Tests" --space-id sp_abc123 --output json
# Create using a specific profile
ax projects create --name "Production Project" --space-id sp_abc123 --profile production
Remove a project by ID:
ax projects delete <project-id>
Arguments:
id (required) - The project IDOptions:
--force, -f - Skip confirmation prompt--profile, -p <name> - Configuration profile to use--verbose, -v - Enable verbose logsExamples:
# Delete with confirmation prompt
ax projects delete proj_abc123
# Delete without confirmation
ax projects delete proj_abc123 --force
# Delete from production
ax projects delete proj_abc123 --profile production
Warning: Deletion is permanent. Always verify the project ID before deleting.
The projects list command uses cursor-based pagination. The response includes a cursor for fetching the next page:
# First page
ax projects list -n 10 --output json
# Use the cursor from the previous response to get the next page
ax projects list -n 10 --cursor <cursor-from-previous-response>
# 1. List all projects
ax projects list --output json | jq '.[] | {id: .id, name: .name}'
# 2. Extract the project ID by name
PROJECT_ID=$(ax projects list --output json | jq -r '.[] | select(.name == "ML Experiments") | .id')
# 3. Get detailed information
ax projects get "$PROJECT_ID"
# 1. Create the project
ax projects create --name "New Experiment" --space-id sp_abc123
# 2. Find the new project ID
PROJECT_ID=$(ax projects list --output json | jq -r '.[] | select(.name == "New Experiment") | .id')
echo "Created project: $PROJECT_ID"
# 3. Verify details
ax projects get "$PROJECT_ID"
# List projects in production
ax projects list --profile production
# Create project in staging
ax projects create --name "Test Project" --space-id sp_staging_123 --profile staging
# Get project details from dev
ax projects get proj_dev_456 --profile dev
# 1. List all projects
ax projects list --output json | jq '.[] | {id: .id, name: .name}'
# 2. Review and identify projects to delete
# 3. Delete old projects
ax projects delete proj_old_001 --force
ax projects delete proj_old_002 --force
Human-readable table with columns for ID, Name, Created, and other metadata.
Structured JSON with full project metadata:
{
"id": "proj_abc123",
"name": "ML Experiments",
"space_id": "sp_xyz789",
"created_at": "2024-01-15T10:30:00Z"
}
Use --output csv or --output parquet for data-processing-friendly formats. You can also pass a file path to write directly to a file:
ax projects list --output projects.csv
ax projects list --output projects.parquet
ax projects listax config showax config show --expandax config initax config showax config show--verbose for more detail: ax projects list --verbosePROJECT_ID=$(ax projects list --output json | jq -r '.[] | select(.name == "My Project") | .id')
ax projects list --output json | jq '.[] | .id'ax projects list --output json | jq '.[] | {id, name}'ax projects get "$PROJECT_ID" to confirm before deletingprod, staging, dev--force in scripts: Skip interactive confirmation with -f when automating/arize-datasets to manage datasets within projectsUse this skill when users want to:
Don't use this skill for:
/arize-datasets instead)/setup-arize-cli instead)