Withings MCP

PassAudited by VirusTotal on May 6, 2026.

Overview

Type: OpenClaw Skill Name: withings-mcp Version: 0.1.1 The skill bundle provides instructions for an AI agent to install and configure the 'withings-mcp-unofficial' package via npx to access Withings health data. The SKILL.md file includes explicit safety guidelines directing the agent not to expose OAuth tokens or private user data stored in ~/.withings-mcp/, and the overall logic is consistent with its stated purpose of providing a Model Context Protocol (MCP) interface for health metrics.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Connecting this skill can allow an MCP client to access private Withings health and wellness records through the authenticated account.

Why it was flagged

The skill requires OAuth-backed access to a Withings account, which is expected for this connector but gives the MCP package delegated access to sensitive account data.

Skill content
Withings signed OAuth tokens stay under ~/.withings-mcp/.
Recommendation

Only authenticate with a Withings account you intend to expose to the MCP client, and do not share or print token files.

What this means

The actual code handling setup, authentication, and data access will come from the npm package resolved at install/runtime.

Why it was flagged

The setup directs users to execute an external npm package without a pinned version, while the provided artifact set contains only documentation and no package source for review.

Skill content
npx -y withings-mcp-unofficial setup
Recommendation

Verify the npm package and linked repository before running setup, and consider pinning a known-good version if supported.

What this means

If the connected MCP server supports mutations or live calls, an agent could affect provider-side or local data unless the user keeps approval boundaries clear.

Why it was flagged

The documentation acknowledges possible live provider calls or writes, but it also gives safer sequencing guidance and calls for explicit consent.

Skill content
Prefer connection_status, manifest, doctor, privacy_audit, or dry-run surfaces before any write or live provider call.
Recommendation

Use status, manifest, privacy audit, and dry-run modes first, and require explicit user approval before any write or live provider action.

What this means

A trusted MCP client may receive health-related data such as body composition, sleep, activity, workouts, and heart records.

Why it was flagged

The skill is intended to expose Withings data through MCP-compatible clients, which is expected but creates a sensitive data boundary between the local connector and any configured agent.

Skill content
helping Claude, Codex, Cursor, Hermes, OpenClaw, or another MCP-compatible client use this project
Recommendation

Configure the MCP server only in clients you trust, and review what data the client can request before using it.