Security Audit: 3 finding(s) in @modelcontextprotocol/server-puppeteer

[starbuck100]

AgentAudit Security Audit Report

Metric	Value
Package	@modelcontextprotocol/server-puppeteer
Version	2025.5.12
Risk Score	11/100
Result	safe
Findings	6 total (0 critical, 0 high, 3 medium, 1 low) — 3 real, 3 by-design

Findings Summary

Severity	Title	File	Line	by_design
Medium	npx -y without version pinning in configuration examples	README.md	113	No
Medium	Unvalidated executablePath in launchOptions	dist/index.js	120	No
Low	Error messages expose internal details	dist/index.js	249	No
High	Arbitrary JavaScript execution via puppeteer_evaluate	dist/index.js	331	Yes (by-design)
Medium	Chrome --no-sandbox in Docker mode	dist/index.js	142	Yes (by-design)
Medium	ALLOW_DANGEROUS env var bypasses argument validation	dist/index.js	124	Yes (by-design)

Key Recommendations

1. Pin version in npx examples — Documentation recommends npx -y @modelcontextprotocol/server-puppeteer without version pinning. Consider @modelcontextprotocol/server-puppeteer@2025.5.12.
2. Validate executablePath — The launchOptions.executablePath parameter is not validated against a known set of Chrome/Chromium binaries. Add it to the dangerous options check or require allowDangerous=true when set.
3. Sanitize error messages — Raw error.message content is returned to MCP clients, potentially leaking internal file paths or system details.

Full Report

View the complete audit report with details, evidence, and remediation guidance:
AgentAudit Report

---

This audit was performed automatically by AgentAudit, the security registry for AI agent packages. If you believe any finding is incorrect, you can dispute it on the platform.

[starbuck100]

Closing this issue — after review, we determined that the severity ratings in this automated audit were overly aggressive and some findings do not represent real security risks. We are recalibrating our detection thresholds. Apologies for the noise.

— AgentAudit Team