AI Security Review
scanned 4d ago · by lpm-firewall-aiNo confirmed malicious attack surface. Risky primitives are tied to user-invoked CMS/Prisma maintenance commands and package-aligned update/Vercel checks, with no lifecycle execution or credential exfiltration found.
Decision evidence
public snapshot- dist/commands/migrate.js interpolates user migration name into execSync command without escaping, but only in explicit `actuate migrate create <name>` use.
- dist/commands/mcp.js writes editor MCP configs containing an API key, but only via explicit `mcp:init` and adds paths to .gitignore.
- package.json has no install/preinstall/postinstall lifecycle hooks.
- dist/index.js only registers Commander subcommands; no import-time network, shell, or file mutation observed.
- dist/commands/db-status.js runs `npx prisma` status/execute checks only after explicit `db:status`.
- dist/utils/database.js dynamically imports project Prisma/cms-core modules for database access, consistent with CLI database features.
- dist/commands/upgrade.js and update-check.js fetch Actuate/npm version metadata and do not execute returned code.
- dist/vercel/client.js only reads Vercel env metadata using user-provided/standard Vercel token.
Source & flagged code
4 flagged · loading sourcePackage source references child process execution.
dist/commands/db-status.jsView on unpkg · L1Package source invokes a package manager install command at runtime.
dist/commands/db-status.jsView on unpkg · L80Package source references dynamic require/import behavior.
dist/utils/database.jsView on unpkg · L21This package version adds a dangerous source file absent from the previous stored version.
dist/commands/doctor.jsView on unpkg