Agent Observability MCP and Skills

Overview

The Datadog MCP Server enables AI agents to access your Agent Observability data through the Model Context Protocol (MCP). The llmobs toolset provides tools for searching and analyzing traces, inspecting span details and content, and evaluating experiment results directly from AI-powered clients like Cursor, Claude Code, or OpenAI Codex.

Setup

Connect an MCP-compatible client to the Datadog MCP Server with the llmobs toolset enabled.

Prerequisites

• A Datadog account with permission to access Agent Observability data.
• An MCP-compatible client (for example, Claude Code, Codex CLI, Cursor, Gemini CLI, or Kiro CLI).

Endpoint

The MCP Server endpoint depends on your Datadog site. Use the Datadog Site selector to display the endpoint for your site. Append ?toolsets=llmobs,core to enable the Agent Observability and core toolsets.

?toolsets=llmobs,core

Connect

Choose remote authentication when possible. Use local binary authentication if your environment blocks the remote OAuth flow.

claude mcp add --transport http datadog-mcp "?toolsets=llmobs,core"

[mcp_servers.datadog]
url = "?toolsets=llmobs,core"

{
 "mcpServers": {
 "datadog": {
 "type": "http",
 "url": "?toolsets=llmobs,core"
 }
 }
}

Authenticate with API keys

The MCP Server uses OAuth 2.0 by default. If OAuth is unavailable, send a Datadog API key and application key as the DD_API_KEY and DD_APPLICATION_KEY HTTP headers:

{
 "mcpServers": {
 "datadog": {
 "type": "http",
 "url": "?toolsets=llmobs,core",
 "headers": {
 "DD_API_KEY": "<YOUR_API_KEY>",
 "DD_APPLICATION_KEY": "<YOUR_APPLICATION_KEY>"
 }
 }
 }
}

For security, scope the API key and application key to a service account with only the required permissions.

Agent skills

Agent skills are prebuilt instruction sets for AI coding agents that automate common Agent Observability workflows. The dd-llmo skill set is available in the Datadog agent-skills repository. It provides six skills for classifying sessions, diagnosing failures, analyzing experiments, generating experiment code with the ddtrace.llmobs SDK, and bootstrapping evaluators against your live production data.

Install

Install the dd-llmo skills with the following command:

npx skills add datadog-labs/agent-skills --skill dd-llmo --full-depth -y

The skills require the llmobs MCP toolset to be connected. If you have not already connected it, run:

claude mcp add --scope user --transport http "datadog-llmo-mcp" \
 'https://mcp.datadoghq.com/v1/mcp?toolsets=llmobs'

Restart Claude Code after running both commands for the skills to appear.

Available skills

Session classification

/llm-obs-session-classify classifies whether user intent was satisfied in a given interaction. It draws from up to three signal sources: Agent Observability traces, RUM behavioral data, and Audit Trail events. The skill returns a yes / partial / no verdict with supporting evidence. Confidence improves with each additional signal source.

/llm-obs-session-classify session_id=<SESSION_ID>
/llm-obs-session-classify trace_id=<TRACE_ID>
/llm-obs-session-classify ml_app=my-chatbot --timeframe now-7d

Trace root cause analysis

/llm-obs-trace-rca diagnoses why an LLM application is producing poor results. It selects an analysis mode based on the strongest available signal (LLM-judge eval verdicts, runtime errors, or structural anomalies) and compiles a structured RCA report. The report includes a failure taxonomy and concrete BEFORE / AFTER fix proposals grounded in trace evidence.

When Claude Code has access to your codebase, the skill can search for the relevant source files and propose diffs inline.

/llm-obs-trace-rca ml_app=my-chatbot
/llm-obs-trace-rca ml_app=my-chatbot eval_name=faithfulness --timeframe now-24h

Evaluator bootstrap

/llm-obs-eval-bootstrap analyzes production traces and proposes a suite of evaluators targeting the observed failure modes. It outputs one of three artifacts: Python BaseEvaluator / LLMJudge classes for offline experiments, a framework-agnostic JSON spec, or online LLM-judge evaluators published directly to Datadog.

/llm-obs-eval-bootstrap ml_app=my-chatbot
/llm-obs-eval-bootstrap ml_app=my-chatbot --publish
/llm-obs-eval-bootstrap ml_app=my-chatbot --data-only

Experiment analyzer

/llm-obs-experiment-analyzer retrieves experiment results and surfaces what changed between a candidate and a baseline: which metrics improved, which regressed, and where the candidate underperformed.

/llm-obs-experiment-analyzer experiment_id=<EXPERIMENT_ID>
/llm-obs-experiment-analyzer experiment_id=<CANDIDATE_ID> baseline_id=<BASELINE_ID>

Generate experiment code with the Python SDK

/llm-obs-experiment-py-bootstrap generates a self-contained Python experiment client that uses the ddtrace.llmobs SDK. The output is either a runnable .py script or a Jupyter .ipynb notebook matching the canonical reference notebook style. The dataset can come from a local JSON or CSV file, an existing Datadog dataset fetched by name, or a built-in inline sample. Every generated experiment is tagged with generated_by=claude-code so you can identify and filter Claude-generated experiments in the LLM Experiments list.

/llm-obs-experiment-py-bootstrap
/llm-obs-experiment-py-bootstrap --dataset ./data/qa.json --format ipynb
/llm-obs-experiment-py-bootstrap --dataset-name <DATASET_NAME> --project-name <PROJECT_NAME>

End-to-end eval pipeline

/llm-obs-eval-pipeline chains session classification, trace RCA, and evaluator bootstrap into a single supervised workflow with user checkpoints between phases. It is the recommended starting point when you have no existing evaluators for an application.

/llm-obs-eval-pipeline my-chatbot
/llm-obs-eval-pipeline my-chatbot --timeframe now-30d --publish

For a complete guide to these skills and a recommended end-to-end workflow, see Analyze LLM Applications with Claude Code Skills.

Use cases

The Agent Observability MCP tools enable AI-assisted workflows for:

• Debugging agent execution: Search for traces by ML app, error status, or custom tags, then examine span hierarchies and content to identify failures.
• Analyzing trace structure: Visualize the full span tree of a trace to understand how agents, LLMs, tools, and retrievals interact.
• Investigating agent loops: Review an agent’s step-by-step execution loop to understand decision-making and tool invocation patterns.
• Evaluating experiments: Get summary statistics for experiment metrics, compare results across dimension segments, and inspect individual events.
• Discovering experiment patterns: Filter and sort experiment events by metric performance to find the best and worst-performing cases.
• Managing evaluators: List, inspect, create, update, and delete evaluator configurations across an ML application or the entire organization.
• Exploring Patterns: List pattern configurations, check run status, and browse the discovered topic hierarchy to understand what users are asking and how traffic is distributed.

Available tools

The llmobs toolset includes the following tools:

Trace and span tools

search_llmobs_spans

Search for spans matching filters or a raw query.

get_llmobs_trace

Get the full structure of a trace as a span hierarchy tree, including span counts by kind, error indicators, and total duration.

get_llmobs_span_details

Get detailed metadata for one or more spans, including timing, error info, LLM details (model, token counts), metrics, and evaluations.

get_llmobs_span_content

Retrieve the actual content of a span field (input, output, messages, documents, or metadata) with optional JSONPath extraction.

find_llmobs_error_spans

Find all error spans in a trace with propagation context, grouped by span kind with error messages and stack traces.

expand_llmobs_spans

Load children of specific spans for progressive tree exploration when get_llmobs_trace returns collapsed nodes.

get_llmobs_agent_loop

Get a chronological view of an agent’s execution loop, showing each step (LLM calls, tool invocations, decisions) in order.

Experiment tools

get_llmobs_experiment_summary

Get a high-level experiment summary with pre-computed statistics for all evaluation metrics. Start here before using other experiment tools.

list_llmobs_experiment_events

List experiment events with filtering by dimension or metric and sorting by metric value.

get_llmobs_experiment_event

Get full details for a single experiment event, including input, output, expected output, all metrics, and dimensions.

get_llmobs_experiment_metric_values

Get statistical analysis for a specific evaluation metric, optionally segmented by a dimension for comparison.

get_llmobs_experiment_dimension_values

Get unique values for a dimension with counts, useful for discovering valid filter and segment values.

Evaluator tools

list_llmobs_evals

List every LLM-judge evaluator configured across all ML applications. Returns each evaluator’s name, ml_app, and enabled status.

list_llmobs_evals_by_ml_app

List all LLM-judge evaluators configured for a specific ML application.

get_llmobs_evaluator

Retrieve an LLM-judge evaluator configuration by name, including its target (ml_app, sampling, filter), LLM provider, and judge prompt template.

create_or_update_llmobs_evaluator

Create or update an LLM-judge evaluator configuration. Targets a specific ML application and optionally a filter or sampling percentage; the judge’s model and prompt template define how it scores each span.

delete_llmobs_evaluator

Delete an LLM-judge evaluator configuration by name.

Patterns tools

list_llmobs_pattern_configs

List all Patterns configurations for the org. Returns each config’s id, name, evp_query, sampling settings, and timestamps. Start here to find a config_id.

get_llmobs_pattern_config

Get the most-recently-modified Patterns configuration for the org.

get_llmobs_pattern_run_status

Get the status and per-activity progress of the most recent Patterns run for a config. Use this to check whether clustering is running, completed, or failed before reading topics.

list_llmobs_pattern_runs

List all completed Patterns runs for a config, newest first. Returns each run’s id, status, timestamps, and the config_snapshot used.

get_llmobs_patterns

Get the topic hierarchy discovered by a Patterns run. Topics are organized into levels, each with a name, description, and point_count. Omit run_id to read the most recent completed run.

get_llmobs_patterns_with_points

Get the topic hierarchy for a run with span IDs inlined on each leaf topic. Set include_metrics=true to also include per-span duration, cost, token counts, and evaluations.

get_llmobs_pattern_points

Get a cursor-paginated page of clustering points (individual spans) assigned to a single topic. Each point includes the span_id, session_id, and a span input preview. Pass next_page_token back as page_token to continue paging.

Recommended workflows

Trace analysis

1. Search: Use search_llmobs_spans to find traces by ML app, status, span kind, or custom tags.
2. Visualize: Use get_llmobs_trace to see the full span hierarchy tree.
3. Inspect: Use get_llmobs_span_details to get metadata, timing, and evaluations for specific spans.
4. Read content: Use get_llmobs_span_content to retrieve the actual I/O, messages, or documents.
5. Debug errors: Use find_llmobs_error_spans to locate all errors in a trace with propagation context.
6. Expand: Use expand_llmobs_spans to load children of collapsed spans for deeper exploration.
7. Agent review: Use get_llmobs_agent_loop to see the step-by-step execution flow of an agent span.

Experiment analysis

1. Summarize: Use get_llmobs_experiment_summary to get overall statistics and discover available metrics and dimensions.
2. Browse events: Use list_llmobs_experiment_events to find events of interest, filtering by dimension or sorting by metric.
3. Inspect events: Use get_llmobs_experiment_event to view full details for a specific event.
4. Analyze metrics: Use get_llmobs_experiment_metric_values to get percentile distributions, true/false rates, or compare across dimension segments.
5. Discover dimensions: Use get_llmobs_experiment_dimension_values to find valid filter and segment values.

Patterns analysis

1. List configs: Use list_llmobs_pattern_configs to find available Patterns configurations and their config_id values.
2. Check run status: Use get_llmobs_pattern_run_status to verify the most recent run is complete.
3. Read topics: Use get_llmobs_patterns to get the full topic hierarchy with names, descriptions, and coherence scores.
4. Inspect spans: Use get_llmobs_patterns_with_points to get topics with span IDs inlined, or get_llmobs_pattern_points to page through the spans of a specific topic.
5. Analyze span content: Use get_llmobs_span_details or get_llmobs_span_content with the span_id values from the previous step to inspect the actual inputs, outputs, and metadata of individual spans within a topic.
6. Browse past runs: Use list_llmobs_pattern_runs to see historical runs and pass a specific run_id to compare topic distributions over time.

Example prompts

After connecting, try prompts like:

• Review error traces for my customer-support-bot app over the past week. Summarize the most common failure patterns, how often they occur, and recommend which ones to fix first.
• Find traces where my agent’s responses were flagged by evaluations as low quality. Look at the inputs and outputs, then suggest specific changes to my system prompt to improve response quality.
• Look at recent agent traces for my app and find cases where the agent looped more than necessary. Analyze the decision-making at each step and suggest how to improve my tool descriptions to reduce unnecessary tool calls.
• A user reported a bad response. Here’s the trace ID: trace-123. Walk me through exactly what happened: what the user asked, what the agent did at each step, and where things went wrong. Suggest a code fix.
• Analyze experiment exp-456 and generate a markdown table of the worst-performing dimensions broken down by evaluation scores. Include any other relevant columns that help me understand where and why performance is degrading.
• Compare experiment exp-123 (baseline) against experiment exp-456. Summarize what improved, what regressed, and by how much. Give me a recommendation on whether the changes are worth shipping.
• Summarize experiment exp-456 and identify the top 5 lowest-scoring events. For each, show the input, output, and which evaluations failed.

Combine with other Datadog tools

The core toolset included in the setup URL gives your AI agent access to additional Datadog tools that pair naturally with Agent Observability analysis.

Export analysis to Datadog Notebooks

The core toolset includes create_datadog_notebook and edit_datadog_notebook, which let your AI agent create Datadog Notebooks directly from analysis results. You can export findings from agent chats into a collaborative, shareable notebook that lives in Datadog alongside your traces and experiments.

Try prompts like:

• Analyze experiment exp-456, identify the worst-performing dimensions, and export a summary report to a Datadog Notebook with a breakdown by evaluation scores.
• Review error traces for my customer-support-bot over the past week and create a Datadog Notebook with the findings, including common failure patterns and recommended fixes.

For custom visualizations that go beyond standard Datadog widgets, like comparison charts or quadrant plots, Notebooks also render Mermaid diagrams natively. Try prompts like:

• Analyze experiment exp-456, compare the accuracy scores across each prompt version, and export the results to a Datadog Notebook that includes a Mermaid bar chart of the average score for each version.
• Analyze experiment exp-456 and export a Datadog Notebook that plots each prompt version on a Mermaid quadrant chart with relevance on one axis and accuracy on the other. Identify which versions are underperforming on both dimensions.

Further reading

Request a personalized demo