beast-agent@2.2.1
Ultra-fast local agent shell for Windows.
AI Security Review
scanned 2mo ago · by lpm-firewall-aiThis is a dangerous-capability package with install-time dependency mutation and permissive runtime agent tools, but the inspected source does not prove concrete malicious behavior. Warn rather than block.
Decision evidence
public snapshot- The install hook deletes Electron runtime files and invokes Electron's installer, including a third-party mirror fallback.
- The AI agent exposes command execution, file writes, and Python execution while its approval gate is disabled by default.
- Command execution launches PowerShell with execution-policy bypass.
- File writes accept normalized absolute paths without a workspace-boundary check.
- The application can persist web watchers that fetch configured URLs periodically.
- No runtime dependency on beast-agent itself is declared.
- The install hook targets the bundled Electron dependency rather than an AI-agent configuration surface.
- No code inspected sends harvested environment variables, credentials, or local files to an attacker-controlled endpoint.
- Remote watcher URLs and command tools are application capabilities, not automatic install-time actions.
Source & flagged code
19 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgThis package version adds a dangerous source file absent from the previous stored version; route for source-aware review.
bin/beast-agent.jsFindings
Affected versions and remediation
This report applies to beast-agent@2.2.1.
See all versions of beast-agent for other recorded verdicts.
- Review the evidence and your use of beast-agent@2.2.1 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.