@vibekiln/cutline-mcp-cli-staging@0.6.0
CLI and MCP servers for Cutline, including SlopBurn: a product quality engineering roguelike RPG for vibecoding workflows.
AI Security Review
scanned 3d ago · by lpm-firewall-aiAn authenticated, user-run `init` command silently sends setup and use events to Cutline's own service, including project-root-derived hashes and the account bearer token; I found no user notice or opt-out in the package. This is undisclosed runtime telemetry, so it warrants a warning rather than a block; no unrelated credential destination was established.
Decision evidence
public snapshot- Authenticated `init` calls installation registration and then reports completion events.
- Registration and event data are sent by POST with the user's Cutline bearer token.
- Registration includes source and host metadata plus IDs hashed from the project root.
- Completion events include the command, account tier, product graph status, and generated rule count.
- The configured telemetry host is Cutline's own service.
- Telemetry is triggered by the explicit `init` command, not by package installation or import.
- The source shows requests to Cutline's service, with no unrelated credential destination identified.
- The project-root-derived identifiers are hashed.
Source & flagged code
8 flaggedThis package version adds a dangerous source file absent from the previous stored version; route for source-aware review.
dist/commands/serve.jsView on unpkgPackage source references child process execution.
dist/commands/serve.jsView on unpkg · L1Findings
Affected versions and remediation
This report applies to @vibekiln/cutline-mcp-cli-staging@0.6.0.
See all versions of @vibekiln/cutline-mcp-cli-staging for other recorded verdicts.
- Review the evidence and your use of @vibekiln/cutline-mcp-cli-staging@0.6.0 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.