registry / @getfrontline/cli / 1.5.1

@getfrontline/cli@1.5.1

Frontline CLI — Public API (agents, workflows, billing, tables, objects) + Max chat/admin REST from your terminal

AI Security Review

scanned 1mo ago · by lpm-firewall-ai

Installing the package automatically modifies a global Claude Code skill directory. This is an unconsented write to a foreign AI-agent control surface.

Static reason

One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review

Trigger

npm installation runs the postinstall lifecycle hook when Claude Code is detected.

Impact

The package can add or update many agent instruction files outside the installed package without an explicit setup command.

Mechanism

Postinstall copies bundled SKILL.md files into the global Claude configuration.

Attack narrative

The npm postinstall hook starts a compiled installer automatically. When it detects Claude Code, it resolves the user-wide Claude configuration directory and copies numerous bundled skill instruction files there, creating directories and overwriting package-owned prior versions. This changes an external AI agent's behavior as a side effect of installing a CLI, rather than after an explicit setup action.

AI rationale

This package has a complete automatic lifecycle chain that writes bundled instructions into the user-wide Claude Code control surface. The foreign, broad agent configuration mutation is concrete and unconsented.

Evidence

package.jsonscripts/runPostinstall.mjsdist/scripts/postinstall.js~/.claude/skills/<skill>/SKILL.md~/.claude/skills/<skill>/.frontline-hash

@getfrontline/cli@1.5.1

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Source & flagged code