luckiest-co@1.0.27
Luckiest for Claude Code: plan, go, finish. Your luckiest.co skills and tribe, inside Claude.
AI Security Review
scanned 25d ago · by lpm-firewall-aiThe source supports a conditional credential exposure risk: a saved service key is sent to an unrestricted environment-configured endpoint. The default endpoint is package-aligned and the installer is explicitly invoked, so the evidence supports a warning rather than a block.
Decision evidence
public snapshot- The installer accepts LUCKIEST_API_URL as the API destination, and sends a saved connection key as a bearer token to that destination when syncing owned skills.
- After the explicit install flow, the installer runs the skill sync; an existing saved key is reused without another prompt.
- The telemetry hook also permits LUCKIEST_API_URL to override its default recipient and attaches the saved skill key when present.
- The global installer removes a duplicate Luckiest command directory only when the marketplace plugin is present and enabled.
- package.json exposes an executable CLI but declares no npm lifecycle scripts, so the sync is not triggered by package installation alone.
- Usage-hook registration is prompted; non-interactive runs decline unless --hook is explicitly provided, and local installs do not register it.
- Telemetry filters out non-Luckiest skills and sends a skill identifier rather than prompt text or tool output.
Source & flagged code
11 flaggedA single source file combines environment access, network access, and code or shell execution; review context before blocking.
bin/install.jsView on unpkg · L187Findings
Affected versions and remediation
This report applies to luckiest-co@1.0.27.
See all versions of luckiest-co for other recorded verdicts.
- Review the evidence and your use of luckiest-co@1.0.27 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.