Loading npm security reports…

@sshaaf/migiq@0.2.3: AI-agent control-surface risk | LPM Firewall

registry / @sshaaf/migiq / 0.2.3

@sshaaf/migiq@0.2.3

AI-powered application migration orchestrator for modernizing legacy applications to cloud-native platforms

AI Security Review

scanned 1mo ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. Installing the package automatically modifies Claude and Cursor agent-control directories in the consumer project or home directory. It also runs a separate local CLI installer before copying package-provided skills and an agent.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

npm postinstall during package installation

Impact

The package can install behavior-bearing skills and an autonomous agent without an explicit setup command.

Mechanism

Automatic external command execution and AI-agent control-surface writes

Policy narrative

When npm installs this package, its postinstall starts install.js. The installer runs rgctl install --skill, then copies eight package skills into both .claude and .cursor and installs AGENT.md as a migrator agent. These are broad behavior-bearing AI-agent control surfaces in either the consuming project or the user home directory, and the mutation occurs without an explicit user setup command.

AI rationale

This is an automatic postinstall mutation of foreign AI-agent control surfaces combined with execution of another CLI installer. Under the isolation policy, that concrete chain requires blocking even though no direct secret theft was found.

Evidence

package.jsoninstall.js.claude/skills/.cursor/skills/.claude/agents/migrator/.cursor/agents/migrator/AGENT.mdmig-rgctl/migiq/mig-plan/mig-prompt-builder/mig-execute/

Source & flagged code

6 flagged · loading source

package.jsonView file

•scripts.postinstall = node install.js

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 1

High

Semantic Analysis Limited

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

package.jsonView on unpkg

•scripts.postinstall = node install.js

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg

install.jsView file

76const skillPaths = [ L77: path.join(cwd, '.claude', 'skills', 'rgctl', 'SKILL.md'), L78: path.join(cwd, '.cursor', 'skills', 'rgctl', 'SKILL.md'), ... L96: // Create destination directory L97: fs.mkdirSync(dest, { recursive: true }); L98: ... L108: } else { L109: fs.copyFileSync(srcPath, destPath); L110: } ... L118: const destDir = path.dirname(dest); L119: fs.mkdirSync(destDir, { recursive: true }); L120: fs.copyFileSync(src, dest);

Critical

Ai Agent Control Hijack

Source creates an unconsented AI-agent control surface through install-time mutation or a default unauthenticated remote skill channel.

install.jsView on unpkg · L76

•Manifest-trigger-reachable source links an external AI-agent control path to a behavior-bearing write operation. #!/usr/bin/env node const fs = require('fs'); const path = require('path'); const os = require('os'); const { execsync } = require('child_process'); const packagejson = require('./package.json'); const version = packagejson.version; function commandexists(cmd) { try { const check = process.platform === 'win32' ? `where ${cmd}` : `which ${cmd}`; execsync(check, { stdio: 'ignore' }); return true; } catch { return false; } } const rgctl_candidate_paths = [ 'rgctl', path.join(os.homedir(), '.local', 'bin', 'rgctl'), '/opt/homebrew/bin/rgctl', '/usr/local/bin/rgctl', ];

High

Trigger Reachable External Ai Agent Control Surface Mutation

Manifest-trigger-reachable source writes behavior-bearing configuration into a user or project AI-agent control surface.

install.jsView on unpkg

mig-plan/evals/aggregate_benchmark.pyView file

•path = mig-plan/evals/aggregate_benchmark.py kind = build_helper sizeBytes = 10547 magicHex = [redacted]

Medium

Ships Build Helper

Package ships non-JavaScript build or shell helper files.

mig-plan/evals/aggregate_benchmark.pyView on unpkg