registry / ccclub / 0.6.16

ccclub@0.6.16

Claude Code and Codex leaderboard among friends for coding agent tokens and costs

AI Security Review

scanned 6d ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. npm postinstall silently mutates Claude Code's settings to add automatic command hooks. It also refreshes an existing macOS LaunchAgent that runs the package every five minutes.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

npm install

Impact

Automatically executes version-pinned package code after Claude sessions and via launchd.

Mechanism

install-time foreign AI-agent hook mutation and persistence refresh

Policy narrative

Installing the package executes postinstall, which rewrites Claude Code SessionEnd and Stop hooks to invoke npx --yes ccclub@0.6.16 sync --silent without an explicit setup command. On macOS, it also reloads an existing LaunchAgent that invokes the same command every five minutes. The runtime sync path maintains these controls and uploads collected usage data to the configured ccclub API.

Rationale

This is concrete, unconsented postinstall mutation of a foreign AI-agent control surface, with persistent automatic execution. The conditional existence check does not make the configuration change user-consented.

Evidence

package.jsonscripts/postinstall.cjsdist/index.js~/.claude/settings.json~/Library/LaunchAgents/dev.ccclub.sync.plist~/.ccclub/sync.log

Network endpoints2

ccclub.devapi.anthropic.com/api/oauth/usage

ccclub@0.6.16

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code