clavue-v1@1.15.5
Clavue v1: lightweight four-slot (main/light/plan/review) multi-model coding CLI on the execution-first engine — overlapped cross-family review, official Clavue cloud preset (api.clavue.com) alongside any Messages-compatible provider, learned context ceil
AI Security Review
scanned 13d ago · by lpm-firewall-aiSource inspection found a CLI package without an install lifecycle hook and did not establish the scanner’s alleged credential exfiltration behavior. The inspected network and subprocess operations align with user-invoked provider and coding CLI functionality.
Decision evidence
public snapshot- package.json defines a CLI binary and no preinstall, install, or postinstall hook.
- The flagged process execution in dist/mao-command.js passes a command and arguments to a runner, consistent with the package’s interactive coding CLI capability.
- The sandbox runtime CLI starts its command-line main function when invoked as a CLI entry point.
- dist/provider-router.js reads Anthropic credentials while constructing an API client and allows an environment-configured base URL, consistent with a configurable provider client; no separate credential theft behavior was established.
- The sandbox HTTP proxy filters destination host and port before forwarding connections.
- Static hints report credential exfiltration, but inspected excerpts establish credential use for an API client and do not show an unrelated recipient or exfiltration sink.
Source & flagged code
11 flagged · loading sourcePackage source references child process execution.
dist/mao-command.jsView on unpkg · L4Package source references shell execution.
node_modules/@anthropic-ai/sandbox-runtime/dist/cli.jsView on unpkg · L44Findings
Affected versions and remediation
This report applies to clavue-v1@1.15.5.
See all versions of clavue-v1 for other recorded verdicts.
- Review the evidence and your use of clavue-v1@1.15.5 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.