by Honeycomb
Updated October 20, 2026
MCP100 verdict
Not currently recommended. First-party evidence supports Shortlist admission; this is not a Recommended-tier security certification.
MCP100 Assessment
Editorial verdict
Reviewed October 19, 2026
Task Fit
Mapped
Specialized workflows
Trust
Status warning
Shortlist evidence is sufficient for identity, setup and task value; authenticated runtime tool-schema measurement and exhaustive security certification remain out of scope.
Maintenance
Date unavailable
No dated repository or package update is currently stored.
Popularity
Signals unavailable
No independently comparable adoption metric is currently stored.
Evidence scope
0 shared-repository stars · not MCP-specific
No MCP-specific package metric is stored
Repository identity: honeycombio/honeycomb-mcp-in-ide-onboarding
Stars observed October 20, 2026 · Downloads observed October 20, 2026
Current Honeycomb repository supports onboarding/skills and connection to the hosted OAuth MCP. It is not the hosted server implementation. The older `honeycombio/honeycomb-mcp` repository is archived and must not be used as current adoption evidence.
Current Honeycomb hosted MCP has no exact public package distribution verified; do not reuse the old archived implementation.
Publisher activity observed October 20, 2026
Current source snapshot
Source relationship: First-party MCP
Reviewed October 19, 2026
Evidence-backed setup
Choose your MCP client. A client connects to the MCP server; the language model is selected inside that client. Only stored, source-backed configurations are shown.
Transport not recorded
Complete client-specific setup, verification and removal are not documented in this reviewed source.
Publisher-documented setup is not the same as MCP100 execution testing. “MCP100 logic checked” means the stored configuration shape was reviewed, not that third-party MCP software was run.
Evidence checked 2026-10-20
Type-appropriate, source-backed setup for this mcp server profile.
Loading optional publisher installation details…
{"url":"https://mcp.honeycomb.io/mcp"}Capability
Honeycomb MCP is a first-party MCP for investigate honeycomb observability data from an mcp-compatible ide or assistant. Authentication/setup: OAuth. The older honeycombio/honeycomb-mcp open-source server was archived in 2026 and must not be treated as the current hosted service or used for its adoption metrics. EU accounts use a regional endpoint.
Current hosted-service onboarding documents representative tools but not a stable complete tools/list count; the archived predecessor repository is not used as tool-count or adoption evidence.
Not measured yet
Estimated token footprint of the MCP tool schemas exposed to the agent.
https://mcp.honeycomb.io/mcp
Reviewed access
Related reviewed MCPs
Query Grafana Cloud metrics, logs, dashboards, and observability context.
Choose this instead if
Choose Grafana Cloud MCP instead if query Grafana Cloud metrics, logs, dashboards, and observability context.
Query Grafana Cloud metrics, logs, dashboards, and observability context.
Investigate Datadog logs, metrics, traces, monitors, dashboards, and security signals.
Choose this instead if
Choose Datadog MCP instead if investigate Datadog logs, metrics, traces, monitors, dashboards, and security signals.
Investigate Datadog logs, metrics, traces, monitors, dashboards, and security signals.
Search and investigate governed Splunk Cloud Platform data through granular MCP tools.
Choose this instead if
Choose MCP Server for Splunk platform instead if security and operations investigation over authorized Splunk searches and platform objects.
Search and investigate governed Splunk Cloud Platform data through granular MCP tools.
Honeycomb is listed as First-party MCP in MCP100’s current evidence.
Current official resources:
From the publisher README
Original client setup and configuration retained from the publisher documentation.
Publisher-documented clients
These names appear in the publisher's installation documentation. MCP100-verified configurations are shown separately above when available.
git clone [url] or download and extract the files"Help me get started with Honeycomb"Your AI assistant will:
Claude Code automatically reads CLAUDE.md files in your working directory.
# Option A: Copy to your project root
cp -r honeycomb-skills/ ~/my-project/
cd ~/my-project
claude # Start Claude Code from this directory
# Option B: Copy to home directory for global access
cp -r honeycomb-skills/ ~/honeycomb-skills/
cd ~/honeycomb-skills
claude
That's it—Claude Code will automatically pick up the CLAUDE.md instructions.
The easiest setup—upload the folder to a Project and the instructions persist across all conversations.
Claude will automatically have access to all the guides and will follow the instructions in CLAUDE.md.
Alternative: Add the contents of CLAUDE.md to ChatGPT's custom instructions (Settings → Personalization → Custom Instructions) for access outside of Projects.
Copy the contents of CLAUDE.md into Cursor's rules file:
cp honeycomb-skills/CLAUDE.md ~/my-project/.cursorrules
cp honeycomb-skills/CLAUDE.md ~/my-project/.windsurfrules
mkdir -p ~/my-project/.github
cp honeycomb-skills/CLAUDE.md ~/my-project/.github/copilot-instructions.md
Point your AI at onboarding/GUIDE.md and ask it to follow the instructions.
Original product, capability, tool, and usage documentation from the MCP publisher. Installation sections appear under Install & Agents.
Learn to investigate your production behavior from inside your code editor using AI + Honeycomb. Instead of learning query syntax and UI navigation, you'll orchestrate investigations by directing an AI assistant that queries Honeycomb on your behalf—using your team's real telemetry data.
This is a different way to learn observability:
New to using AI for technical work? That's okay! This skills pack guides you through it. Think of the AI as a knowledgeable teammate who knows Honeycomb deeply and can run queries for you. You just need to tell it what you're trying to figure out.
See the Honeycomb MCP configuration guide for full setup instructions across all supported tools.
Easiest setup—the onboarding guide will do this for you automatically!
When you start onboarding, Claude will detect if the MCP is missing and set it up for you. Just start with "Help me get started with Honeycomb" and follow the prompts.
Manual setup (if preferred):
claude mcp add honeycomb --transport http https://mcp.honeycomb.io/mcp
# Exit and restart Claude Code
# Run: /mcp to authenticate
For EU region, use https://mcp.eu1.honeycomb.io/mcp instead.
See the configuration guide for tool-specific instructions including Claude Desktop, VS Code, Amazon Q Developer, and more.
Authentication happens via OAuth when you first connect — no API keys needed.
After setting up the MCP, verify everything works before starting the learning paths. Copy and paste these prompts to your AI assistant:
Confirm the Honeycomb MCP is connected and authenticated. List the Honeycomb MCP tools you have access to (names only). If not connected, tell me exactly what to do next.
What should happen: The AI should list MCP tools like get_workspace_context, run_query, get_trace, etc.
If it fails: The AI will guide you through setup or authentication. Follow its instructions, then try this prompt again.
Using Honeycomb MCP, verify you can access my workspace data. Identify the environment and dataset you will use (ask me at most 2 questions if needed). Then run the smallest safe query you can over the last 60 minutes that proves data is present. Show: time window, filters, and a Honeycomb link to the result.
What should happen:
If it fails:
/mcp to re-authenticateIf both prompts succeeded, you're all set. Now start your learning journey:
Help me get started with Honeycomb
The AI will ask what you're trying to do and guide you from there.
This skills pack works differently than traditional tutorials:
Traditional approach: Learn the query language → Learn the UI → Try it on your data → Debug production issues
This approach: Start with what you're trying to do → AI runs queries for you → Learn by seeing real results → Build understanding through practice
When you start, the AI will ask what you're trying to do. Choose based on your current need:
You have a production issue to investigate right now. Learn Honeycomb by walking through a real debugging workflow:
You want to understand what telemetry exists and build a mental model of your services:
You want to understand your service's health and what "good" looks like:
All paths are equal—start with whichever matches what you're trying to do right now. You can come back and try other paths later.
New to Honeycomb? The AI teaches concepts as you encounter them, using your real data as examples. It tracks what you've learned so it never repeats explanations.
Already experienced? Mark yourself as completed: true in onboarding/progress.yaml and the AI will stay concise and out of the way—just helping when you need it.
honeycomb-skills/
├── CLAUDE.md # AI instructions (copy to your tool's config)
├── onboarding/
│ ├── GUIDE.md # Interactive onboarding—learns your experience level
│ ├── setup-mcp.md # Automated MCP setup guide (OAuth-based)
│ ├── progress.yaml # Tracks what you've learned (auto-updated)
│ └── my-context.yaml # Your role and preferences (optional)
├── investigation/
│ └── GUIDE.md # Expert debugging heuristics and techniques
├── slo-basics/
│ └── GUIDE.md # Understanding SLOs, SLIs, and error budgets
└── shared/
└── honeycomb-concepts.md # Reference glossary of Honeycomb terminology
Edit onboarding/my-context.yaml to set your preferences:
role: product_engineer # Your role (product_engineer, sre, data, manager)
experience_with_observability: none # none, some, experienced
primary_services: [checkout, payments] # Services you work on most
explanation_level: detailed # detailed, normal, minimal
You'll know you've successfully learned a path when you can:
Debugging path:
Exploring path:
Reliability path:
The ultimate goal: You should feel comfortable using your AI assistant to investigate production issues on your own—orchestrating queries, interpreting results, and building hypotheses without needing to remember query syntax or UI navigation.
Found something confusing? Want to add a skill? Open an issue or submit a PR.
You can also share feedback directly from your AI assistant — type /feedback at any time and it will walk you through a short survey.