Where this fits

Your agent is on-call. A production alert fires, and you need it to dig into Datadog—pull logs, check monitor status, and correlate metrics—without waiting for human input. The datadog skill gives your agent read-only access to Datadog's API, turning it into a first responder that can investigate incidents and report back with actionable context.

Why agents benefit

• Direct API access – No need to wrap Datadog's REST calls; the skill provides a CLI that handles authentication and endpoint logic.
• Log search and filtering – Agents can run targeted queries like service:api error over a time range, isolating the root cause faster.
• Monitor and alert history – Agents can check current status and recent transitions, helping you understand if a problem is new or recurring.
• Metric and APM queries – Pull numeric trends and trace data to correlate with symptoms, giving context beyond a single log entry.
• Host and infrastructure insights – Agents can verify resource usage or list affected hosts during an outage, rounding out the investigation.

Practical scenarios

Incident triage with log correlation

An alert triggers for high error rates. The agent runs logs search --query "service:api status:500" --from 15m to gather error samples, then queries a metric to confirm the spike. It summarizes the findings, including the top error message and affected endpoints, so you can jump straight to mitigation.

Pre-deployment health check

Before promoting a service, the agent checks monitor status and recent alert history for the target environment. It pulls APM traces to verify latency is within SLOs and lists host status to catch capacity issues—all from a single prompt.

Proactive anomaly detection

Your agent runs a scheduled task that queries metric trends for CPU and memory on key hosts. When a deviation is spotted, it fetches recent logs and tags the trend as a potential issue, flagging the relevant service owner before the problem escalates.

Add it to your agent workflow

Install the skill with:

npx skills add greedychipmunk/agent-skills --skill datadog

Set the required environment variables, then call the CLI directly from your agent's tooling:

export DD_API_KEY="your-api-key"
export DD_APP_KEY="your-application-key"
export DD_SITE="datadoghq.com"

Example: search logs from the last hour
npx tsx scripts/datadog.ts logs search --query "service:api error" --from 1h

The skill is read-only, so you can safely delegate investigative tasks without worrying about accidental mutations.

Compare before adopting

• Quality signals – Check the repo's star count (currently 14) and recency of commits. A single maintainer may mean slower updates, so verify the API changes align with your Datadog version.
• Maintenance freshness – Look for issues/PRs and whether the author responds. A stale skill could break if Datadog's API shifts.
• Alternatives – Consider writing your own wrapper if you need custom endpoints or tighter security controls. This skill is a good starting point but may not cover every Datadog feature.
• Workflow fit – Ensure the CLI's output format is parseable by your agent. It uses tsx, so Node.js is a prerequisite—check your runtime.

Why it is worth tracking

This skill fills a gap for agents that need deep observability access without scaffolding custom code. The low star count suggests early adoption, but the clean API surface (logs, metrics, monitors, APM) and read-only design make it a practical addition to any production-focused agent. Keep an eye on its GitHub for updates—if the maintainer keeps pace with Datadog's API, it becomes a strong candidate for your stack. You can also explore more skills and share feedback at OpenAgentSkill.