trace-mcp@3.36.3
Framework-aware code intelligence MCP server — 88 framework integrations, 81 languages, 72.7% fewer input tokens to review a pull request, comprehension at parity
AI Security Review
scanned 12d ago · by lpm-firewall-aiSource inspection found package-specific installation setup and separate developer benchmark utilities, including one that reads local Claude session logs when invoked. The inspected locations provide no evidence of credential theft, unconsented agent-control mutation, or other concrete attack behavior.
Decision evidence
public snapshot- package.json limits postinstall to three named setup scripts; it does not run the flagged benchmark utilities.
- The two bundled files define paths for Claude and other agent data; the inspected locations do not show configuration writes.
- The name-token benchmark reads repository documentation, while the state replay benchmark scans local Claude session logs when run.
- The remaining cited scripts describe screenshot capture, upgrade and wire checks, documentation generation, and performance or response-token benchmarks.
Source & flagged code
26 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgFindings
Affected versions and remediation
This report applies to trace-mcp@3.36.3.
See all versions of trace-mcp for other recorded verdicts.
- Review the evidence and your use of trace-mcp@3.36.3 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.