What is connect-metrx?

A setup guide for connecting Metrx, a service that reviews AI-agent work, to your company's agents with an API key. It explains where to create the key, how to set it for Claude Code, and how to check the connection.

Use it when switching from Metrx's synthetic sample data to your organization's data, or when a Metrx tool reports that authentication is required.

It removes the guesswork from authentication and avoids exposing the API key in chat. It also explains why Claude Code must be restarted before it can use the key.

connect-metrx — Skill for Claude Code by metrxbots/mcp-server

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add metrxbots/mcp-server --skill connect-metrx

Clone the repo

git clone --depth 1 https://github.com/metrxbots/mcp-server

Made for: Claude Code.

Or install metrx-desk, the plugin that ships this one along with the rest of its 5 skills, 1 MCP server.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/metrxbots/mcp-server/connect-metrx/github.svg)](https://agentmods.dev/skills/metrxbots/mcp-server/connect-metrx)

Your own site

<a href="https://agentmods.dev/skills/metrxbots/mcp-server/connect-metrx"><img src="https://agentmods.dev/badge/skills/metrxbots/mcp-server/connect-metrx/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/metrxbots/mcp-server/connect-metrx"><img src="https://agentmods.dev/badge/skills/metrxbots/mcp-server/connect-metrx.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 66 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 713 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00066	$0.00713
Opus 5.5	$0.00026	$0.00285
Sonnet 5.5	$0.00013	$0.00143
Haiku 4.5	$0.00007	$0.00071

Measured yesterday against content hash 36c603698ba4, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

connect-metrx scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

packages/metrx-desk-plugin/skills/connect-metrx/SKILL.md · 59 lines

How it starts

The opening of the file, as written. The whole thing — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Connect your own agents

The desk reads your own organization through a Metrx API key. Without one, the same tools answer on the Northwind sample (synthetic demo data).

Steps

Ask whether the person already has a Metrx account. If not, the sign-up is at https://metrxbot.com/sign-up. Do not create the account or sign in for them.
Creating the key: in the Metrx dashboard open Settings → Security and create an API key. For the review and the queue the key needs the read scope; to record decisions it needs write; to publish a job contract it needs admin. Keys start with sk_live_. Tell them to copy it once; Metrx shows it once.
Setting the key for this plugin. The plugin's server entry reads METRX_API_KEY from the environment Claude Code was started in. Tell them to add it to their shell profile (or the environment of the process that launches Claude Code), for example:
```
export METRX_API_KEY=sk_live_...
```
Never ask them to paste the key into the chat, and never echo a key back. If they do paste one, do not repeat it.
Restart Claude Code. MCP servers read their environment at start, so the key is not seen until then.
Verify: call metrx_get_workforce_review with { "org": "me" }.
- auth_required: the key is not reaching the server. Check the variable name, the restart, and that the key starts with sk_live_.
- insufficient_scope: the key lacks read. Create one with it.
- A document with state other than ok: relay its state_message. no_events means no agent has sent anything yet; no_outcomes means events arrive but no outcome has been confirmed; timeout means part of the read missed its deadline and the next read fills it in. Each message says what to do next.
- A document with state: "ok": relay it as /metrx-desk:sample-review does, labels included, and say the review page for their org is in the dashboard.
Sending the first events: agents connect through the Python SDK, the TypeScript SDK or OpenTelemetry; the snippets are at https://metrxbot.com/agent-quickstart. Confirmed outcomes come from the SDK's outcome call or from a review signal; without them the cost per confirmed outcome stays "insufficient data", which is the honest state.

Read the full file on GitHub · 59 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

yesterday First seen · 59 lines · 66 tokens per session scan A 36c603698ba4

Subscribe to this mod's changes

connect-metrx is a skill published in the GitHub repository metrxbots/mcp-server (2 stars, last pushed 2d ago), licensed MIT. It adds 66 tokens to every session and 713 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-11-10.

Analytics and metrics Cost Optimization LLM ai-agent finops mcp mcp-server

connect-metrx next to 6 skills from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
connect-metrxmetrxbots/mcp-server	Skill	2	2d ago	MIT	A	66	Claude Code
lynkrFast-Editor/Lynkr	Skill	640	5d ago	Apache-2.0	A	50	Claude Code, Codex
gentech-blockrunBlockRunAI/blockrun-mcp	Skill	390	yesterday	MIT	A	46	Claude Code, Codex
portainer-mcp-hygieneportainer/portainer-mcp	Skill	242	19d ago	MIT	A	315	Claude Code, Codex
hmem-releaseBumblebiber/hmem	Skill	24	10d ago	MIT	A	50	Claude Code
lookup-topicSmart-AI-Memory/attune-ai	Skill	11	2d ago	Apache-2.0	A	46	Claude Code
api-simulationAryan-Jhaveri/mcp-statcan	Skill	8	4d ago	MIT	A	210	Claude Code, Codex