@chemx/starter-kit@26.9.20-408
Chemical X Protocol: Private drop-in architecture starter kit and capsule generator
AI Security Review
scanned 11d ago · by lpm-firewall-aiThe declared postinstall reaches code that writes behavior-bearing MCP configuration into consumer agent surfaces without a consent guard. This meets the install-time AI-agent control-surface mutation policy.
Decision evidence
public snapshot- The package postinstall hook runs cli/postinstall.js.
- The lifecycle handler targets the consumer directory and automatically invokes installAllMcpConfigs without a consent check.
- The invoked installer writes MCP server configuration into the consumer's Cursor and VS Code project configuration.
- When a Gemini config directory exists, the installer also writes MCP configuration to the user's global Antigravity config.
- The configured server launches this package's MCP command, creating an agent integration during npm installation.
- The installer merges the named Chemical X server entry into existing MCP configuration.
- The global Antigravity write is conditional on an existing config directory.
Source & flagged code
11 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 unpkgPackage contains a critical-looking secret pattern.
cli/errors/errors.spec.jsView on unpkg · L27Findings
Affected versions and remediation
This report applies to @chemx/starter-kit@26.9.20-408.
See all versions of @chemx/starter-kit for other recorded verdicts.
- Avoid installing @chemx/starter-kit@26.9.20-408. 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.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.