Pup CLI is a comprehensive, AI-agent-ready command-line interface that gives AI agents access to Datadog’s observability platform. It exposes Datadog’s API surface for use in AI agent workflows and automated pipelines.
Key features:
Self-discoverable commands: Commands are structured so agents can navigate them without external documentation.
Structured output: Responses are available in JSON and YAML for reliable parsing.
Scoped authentication: OAuth2 and PKCE provide scoped access without long-lived API keys.
Broad product coverage: Pup supports monitors, logs, metrics, RUM, security, and more.
This page covers the Pup's core features. See the Pup repository documentation for the full list of features and commands.
Installation
Homebrew (macOS/Linux)
brew tap datadog-labs/pack
brew install datadog-labs/pack/pup
# Log in to Datadogpup auth login
# List monitors filtered by tagpup monitors list --tags="team:api-platform"# Search logs for errors in the last hourpup logs search --query="status:error" --from="1h"# Query CPU metricspup metrics query --query="avg:system.cpu.user{*}" --from="1h"# Get dashboard detailspup dashboards get <DASHBOARD_ID>
# Delete a dashboardpup dashboards delete <DASHBOARD_ID> --yes
Supported product areas
Pup covers most major Datadog product surfaces. See the command reference for the canonical product-specific command list. You can also run pup --help (or pup agent schema for machine-readable output) for the live list of commands as built.
When Pup is invoked by an AI coding agent, it automatically switches to agent mode, which returns structured JSON responses optimized for machine consumption. Responses include metadata, error details, and hints. Agent mode also auto-approves confirmation prompts.
Agent mode is auto-detected for supported coding agents when their environment variable is set. You can also enable it explicitly with the --agent flag or by setting FORCE_AGENT_MODE=1.
Additional features
Pup includes additional features that can be used in AI agent workflows—follow the links below for more information:
Runbooks: pup runbooks is a local execution engine for YAML-defined operational procedures, encoding multi-step tasks using pup, shell, HTTP, and Datadog Workflow steps.
Agent skills: Pup ships skills and domain agents embedded in the binary, installable to any AI coding assistant with pup skills install.
ACP server: pup acp serve runs a local AI agent server that connects coding tools to Datadog Bits AI through ACP and OpenAI-compatible protocols.
Authentication
Pup supports OAuth2 and API key authentication methods. OAuth2 is preferred; run pup auth login to authenticate through your browser. If OAuth2 is not available, Pup falls back to API keys (DD_API_KEY and DD_APP_KEY). See the authentication documentation for details.
Global flags
Flag
Description
-o, --output
Output format (json, table, yaml). Default: json
-y, --yes
Skip confirmation prompts for destructive operations
--agent
Enable agent mode
--no-agent
Disable agent mode
--read-only
Block all write operations (create, update, delete)
--org <org>
Use a named org profile for multi-account workflows (run pup auth login --org to set up)