ddtoolsets

작성자: datadog-labs

Manages toolsets for the Datadog MCP server `plugin:datadog:mcp`. Use when the user wants to view, enable, or disable toolsets that control which tools are…

npx skills add https://github.com/datadog-labs/claude-code-plugin --skill ddtoolsets

Datadog MCP Server

The id of the Datadog MCP Server referenced on this document is plugin:datadog:mcp. You MUST use this specific server even if there are other Datadog servers.

Shared reference

Read references/mcp-settings.md before proceeding. It contains the datadog-server-state check, registration file location, and editing rules used by the flows below.

Entry flow

Check the datadog-server-state (see mcp-settings.md). Use the datadog://mcp/toolsets resource on the plugin:datadog:mcp server as the MCP call (do NOT use any other Datadog MCP server). Do not output anything until the datadog-server-state and resource content are available, and proceed based on the results:

  • datadog-server-state=working AND valid content — without any preamble, go to the Toolsets Flow.
  • datadog-server-state=not-setup — without any preamble, tell the user the plugin is not set up and instruct them to run /ddsetup, and stop.
  • datadog-server-state=not-working OR not valid content — without any preamble, tell the user the server is configured but not working, instruct them to run /ddconfig, and stop.

When communicating with the user below, describe the server state and actions in plain language. Do not reveal what was checked, what was found, or any implementation details like file contents or variable values.

Toolsets Flow

A toolset is a named group of related tools for a specific Datadog feature. Enabling a toolset makes its tools available; disabling it removes them.

How toolset defaults work

The DD_MCP_TOOLSETS default value in the registration file controls which toolsets are active. It has two states:

  • Empty (${DD_MCP_TOOLSETS:-}) — the server decides which toolsets to enable. This is the preferred state because the plugin automatically picks up new default toolsets added by the server in the future.
  • Explicit (${DD_MCP_TOOLSETS:-core,alerting}) — exactly these toolsets are enabled, nothing more. The server's defaults are ignored. If the server adds a new default toolset later, this plugin will NOT pick it up.

The order of toolsets in the comma-separated list is not meaningful. core,alerting and alerting,core are equivalent. When comparing lists (e.g. to check if the result matches the defaults), compare as sets, not strings.

When computing changes, always prefer empty over an explicit list that happens to match the current defaults. See the editing rule in mcp-settings.md for how to set an empty default value.

1. Gather toolset information

Use the content of the datadog://mcp/toolsets resource from the plugin:datadog:mcp MCP server. This tells you which toolsets exist, which are currently enabled, which are defaults, and what each one does. Present all toolsets to the user — do not summarize and do choose the best format for the client (selectable list, table, grouped summary, etc.). Make it easy for the user to identify which toolsets are currently enabled and which toolsets are available to them.

Also read the current DD_MCP_TOOLSETS default value from the registration file. If it is empty, the user is currently using server defaults. If it has an explicit list, those are the manually selected toolsets.

Any toolset name in the registration file that does not appear in the datadog://mcp/toolsets resource is unknown — ignore it when presenting to the user and silently drop it when writing the updated list.

2. Understand the user's intent

The user may want to:

  • Add more toolsets to the currently enabled list
  • Remove toolsets from the currently enabled list
  • Replace the entire list with a specific set of toolsets

Understand the user's intent from their response. Ask for clarification if ambiguous.

Important: If the current default value is empty (server defaults) and the user wants to add a toolset, you need to know what the defaults ARE so you can build the full list. Use the default information from the datadog://mcp/toolsets resource.

3. Compute the new toolset list

Apply the user's changes to produce a new comma-separated value for DD_MCP_TOOLSETS:

  • If the resulting list matches the default toolsets exactly → use an empty string (revert to server defaults).
  • If the user wants to revert to defaults (e.g. "reset", "use defaults") → use an empty string.
  • If all toolsets would be removed → use an empty string and warn the user that the server's default toolsets will be used instead.
  • If the resulting explicit list does not include core → warn the user before applying. The core toolset provides essential Datadog functionality and most workflows depend on it. Only proceed without core if the user explicitly confirms.
  • Otherwise → use the explicit comma-separated list.

4. Apply the change

Edit DD_MCP_TOOLSETS in the registration file following the editing rule in mcp-settings.md.

Example — adding alerting when currently using server defaults (assuming core and synthetics are defaults):

${DD_MCP_TOOLSETS:-}  →  ${DD_MCP_TOOLSETS:-core,synthetics,alerting}

Example — reverting to server defaults:

${DD_MCP_TOOLSETS:-core,alerting}  →  ${DD_MCP_TOOLSETS:-}

Then silently write the new DD_MCP_TOOLSETS value to ${CLAUDE_PLUGIN_DATA}/toolsets (plain text, one line — write an empty file if reverting to server defaults).

5. Confirm

Tell the user the toolsets have been updated including which toolsets are now enabled, and that they need to follow these steps:

  1. Run the command /reload-plugins
  2. Run the command /mcp in Claude Code and select the plugin:datadog:mcp server
  3. Select the authentication option

datadog-labs의 다른 스킬

dd-audit
datadog-labs
감사 추적 조사 - 누가 무엇을 변경했는지, 키 손상, 비용 급증 근본 원인, 규정 준수 증거(SOC 2/PCI), AI 활동 감사.
official
agent-install
datadog-labs
Datadog Operator를 사용하여 Kubernetes에 Datadog Agent를 설치합니다 — Single Step Instrumentation(SSI)을 활성화하기 전에 필요하며, 이는 자동으로…
official
agent-observability-auto-experiment
datadog-labs
실제 Datadog LLM-Obs 데이터를 대상으로 반복적 코드 개선 힐클라임을 로컬에서 Claude Code를 에이전트로 사용하여 실행합니다. 기준 평가를 설정하고, 하나의…
official
agent-observability-eval-bootstrap
datadog-labs
프로덕션 트레이스에서 평가자를 부트스트랩합니다 — 기본적으로 온라인 LLM-판정 평가자를 제안하고, 확인 후 Datadog에 비활성화된 초안으로 생성합니다…
official
agent-observability-eval-pipeline
datadog-labs
계측된 ml_app을 위한 엔드투엔드 에이전트 관측성 파이프라인 — 프로덕션 트레이스를 분류하고, 실패의 근본 원인을 분석하며, 평가기를 부트스트랩한 다음, (선택적으로)…
official
agent-observability-experiment-analyzer
datadog-labs
LLM 실험 결과를 분석합니다. 단일 또는 비교 실험, 탐색적 또는 Q&A 모드를 처리합니다. 사용자가 "실험 분석", "비교…"라고 말할 때 사용하세요.
official
agent-observability-replay-trace
datadog-labs
개발자가 마음에 들지 않는 출력을 생성한 특정 Agent Observability / LLM Obs 트레이스 하나를 반복 작업하고자 할 때 사용합니다 — 해당 트레이스를 다시 실행하여…
official
agent-observability-trace-rca
datadog-labs
프로덕션 LLM 트레이스에 대한 근본 원인 분석. LLM 애플리케이션이 실패하는 이유를 진단하며, 평가 판정, 런타임 오류 또는 구조적 문제를 기반으로 작동합니다…
official