registry / 3a-factory / 2.2.1

3a-factory@2.2.1

Automated-dev-team AI workflow template for Claude, Gemini, and/or Cursor (selective agent install)

AI Security Review

scanned 3h ago · by lpm-firewall-ai

LPM blocks this version under the AI-agent control-surface policy. npm postinstall automatically mutates the consuming project’s AI-agent control files. With no agent argument it targets all supported agent platforms.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

npm installation of 3a-factory

Impact

Unconsented broad AI-agent control-surface modification; installed instructions can govern future agent behavior in the project.

Mechanism

postinstall copies bundled AI workflow rules, commands, and skills into INIT_CWD

Policy narrative

Installing the package invokes its installer without requiring an explicit user command. The installer selects all supported agents by default and writes bundled instructions, skills, commands, and an always-applied Cursor rule into the consuming project. This is an unconsented postinstall mutation of foreign AI-agent control surfaces.

Rationale

The source establishes the policy-defined blocking condition: automatic lifecycle mutation of broad, foreign AI-agent configuration surfaces. No separate exfiltration or remote execution chain was found.

Evidence

package.jsondist/install.jsdist/bundle.jsonAGENTS.mdWORKFLOW.md.claude/skills/**.claude/commands/**.gemini/skills/**.gemini/commands/**.cursor/skills/**.cursor/rules/ai-workflow.mdc

3a-factory@2.2.1

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code