registry / @macintacos/caret / 0.6.0

@macintacos/caret@0.6.0

A Claude Code plugin that replaces the terminal plan-approval prompt with a local web UI for inline review, annotation, and approval.

AI Security Review

scanned 5h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. This package is a local plan-review extension for Claude Code and OpenCode. When enabled or explicitly installed, it intercepts plan-review flow, launches its local review daemon, and can edit OpenCode’s package-owned plugin configuration.

Static reason

No blocking static signals were detected.; previous stored version diff introduced dangerous source

Trigger

Claude Code hook activation, OpenCode plugin load, or explicit `caret install --target opencode|claude`.

Impact

Can gate or redirect plan approval for the agent environment where the user enables it; no unconsented npm install-time mutation was found.

Mechanism

AI-agent plan-review hook and local UI bridge

Rationale

Source inspection found no npm lifecycle execution, exfiltration, payload retrieval, or destructive behavior. Because this is a first-party agent extension that can alter planning approval behavior when enabled, retain a warning rather than mark clean.

Evidence

package.jsonbin/carethooks/hooks.jsondist/cli.jsopencode/caret.plugin.tsopencode/commands/discovery.mdopencode/commands/demo.mdopencode/commands/debug.md

Network endpoints1

api.github.com/repos/macintacos/caret/releases/latest

@macintacos/caret@0.6.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code