savant-code@0.0.28
Multi-Agent AI Coding Assistant with ECHO Protocol
AI Security Review
scanned 3d ago · by lpm-firewall-aiThe flagged process and network behavior implements the CLI’s binary bootstrap and update flow, with no install hook or import-time launch. Updates are staged for user approval, and telemetry is gated on an environment-provided API key and host.
Decision evidence
public snapshot- The manifest defines only prepack and postpack scripts; neither is an install lifecycle hook.
- The package exports a launcher, but index.js calls launcher.main() only when run as the CLI, not when imported.
- The launcher checks the npm registry for the latest version and downloads release archives from GitHub by default.
- A staged update is installed only after an interactive user approves the prompt.
- PostHog failure reporting returns without sending unless both an API key and host are supplied through environment variables.
- Running the CLI without an installed binary downloads and executes a platform-specific binary, which is a consequential capability but matches the package’s coding-assistant purpose.
- When PostHog reporting is configured, its payload includes the home directory path and update error details.
Source & flagged code
6 flaggedThis package version adds a dangerous source file absent from the previous stored version; route for source-aware review.
launcher.jsView on unpkgFindings
Affected versions and remediation
This report applies to savant-code@0.0.28.
See all versions of savant-code for other recorded verdicts.
- Review the evidence and your use of savant-code@0.0.28 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.