ai-usage-meter@0.1.2
Show Claude, ChatGPT/Codex, and Gemini activity and usage on an AI Meter display.
AI Security Review
scanned 11d ago · by lpm-firewall-aiThe automatic global-install lifecycle hook makes broad AI-agent configuration changes across three providers and installs recurring background behavior. This meets the policy for blocking unconsented install-time mutation of a broad AI-agent control surface.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
Installing the package globally through npm triggers postinstall.
Impact
Agent settings and background behavior are changed during installation without provider-specific consent.
Mechanism
The hook runs setup for all providers, adds agent hooks and status-line configuration, and installs a recurring Codex LaunchAgent without a separate setup command.
Policy narrative
A global npm installation invokes postinstall, which configures Claude, Codex, and Gemini automatically. The setup writes hooks into their settings and installs a recurring Codex LaunchAgent. These are broad agent-control changes triggered during package installation, without requiring a separate setup command.
AI rationale
The automatic global-install lifecycle hook makes broad AI-agent configuration changes across three providers and installs recurring background behavior. This meets the policy for blocking unconsented install-time mutation of a broad AI-agent control surface.
Evidence
package.jsonsetup/postinstall.mjssetup/setup.mjs~/.claude/settings.json~/.codex/hooks.json~/.gemini/settings.json~/Library/LaunchAgents/com.ai-usage-meter.codex-sync.plist
Decision evidence
public snapshotAI identified this as a dangerous AI-agent capability at 98.0% confidence with low false-positive risk.
Evidence for policy risk
- The postinstall hook runs automatically and only skips setup when the package is not installed globally; global installation starts setup for all integrations.
- Setup writes hooks and status-line configuration into Claude, Codex, and Gemini settings, and creates a Codex LaunchAgent for recurring background sync.
Evidence against
- Local installations exit the postinstall hook before configuration begins.
- Setup preserves existing settings and backs them up before writing changes.
Behavioral surface
ChildProcessEnvironmentVarsFilesystem
HighEntropyStringsUrlStrings
Source & flagged code
6 flagged · loading sourcepackage.jsonView file
•scripts.postinstall = node setup/postinstall.mjs
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg37"postinstall": "node setup/postinstall.mjs"
},
"engines": {
High
Suspicious Lifecycle Evidence
The postinstall hook runs automatically and only skips setup when the package is not installed globally; global installation starts setup for all integrations.
package.jsonView on unpkg · L37•scripts.postinstall = node setup/postinstall.mjs
Medium
Ambiguous Install Lifecycle Script
Install-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgsetup/postinstall.mjsView file
Findings
1 Critical3 High4 Medium4 Low
CriticalAi Agent Control Hijacksetup/postinstall.mjs
HighInstall Time Lifecycle Scriptspackage.json
HighSuspicious Lifecycle Evidencepackage.json
HighSuspicious Lifecycle Evidencesetup/postinstall.mjs
MediumAmbiguous Install Lifecycle Scriptpackage.json
MediumEnvironment Vars
MediumShips Native Binaryvendor/picotool/darwin-arm64/libusb-1.0.0.dylib
MediumStructural Risk Force Deep Review
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to ai-usage-meter@0.1.2.
See all versions of ai-usage-meter for other recorded verdicts.
- Avoid installing ai-usage-meter@0.1.2. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.