registry / @einsia/agent-git / 0.2.22

@einsia/agent-git@0.2.22

Version control for agent sessions — the agit CLI plus skills/hooks/MCP wiring for Claude Code, Codex, OpenCode and Cursor. The AgentGit hub deploys separately.

AI Security Review

scanned 21h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. Installation automatically invokes the package's own setup command, which the package describes as configuring agent integrations. This is a guarded first-party agent setup lifecycle risk; no malicious effect was established from inspected source.

Static reason

One or more suspicious static signals were detected.

Trigger

Installing the package with lifecycle scripts enabled and without AGIT_SKIP_SETUP set.

Impact

The setup command may configure agent integrations; inspected source does not establish harmful or foreign control-surface changes.

Mechanism

The postinstall script resolves the packaged binary and invokes it with the setup argument.

AI rationale

Installation automatically runs a first-party agent setup command, which fits the warning category for package-owned agent extension setup. Inspected source does not establish a concrete attack or justify a publish block.

Decision evidence

public snapshot

AI called this Suspicious at 84.0% confidence as Dangerous Capability with medium false-positive risk.

Evidence for warning

The manifest runs a postinstall script during installation.
The postinstall script invokes the installed CLI's setup command unless AGIT_SKIP_SETUP is truthy.
The package describes setup as wiring skills, hooks, MCP, and AGENTS.md for its own CLI.

@einsia/agent-git@0.2.22

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Source & flagged code