aims-workstation
AIMS 工作站:本地智能体工作站——ARM 助手本地化(下载/离线/升级)、工作台对话式定制、多工作区隔离、分级自主与审计
This report shows the evidence for aims-workstation@0.1.27. 3 of 24 known versions are blocked or flagged. The newest version, 0.1.29, is flagged. See all versions.
AI Security Review
scanned 6d ago · by lpm-firewall-aiThe package routes configured LLM credentials to a fixed vendor gateway over plaintext HTTP by default, with TLS verification disabled. This creates a concrete credential disclosure risk when users configure their key without also replacing the default endpoint.
Decision evidence
public snapshot- The default LLM endpoint is the fixed plaintext URL http://bifrost.infra.dev.aimstek.cn/v1.
- The LLM provider resolves the configured or environment API key and uses it with that endpoint for model requests.
- TLS certificate verification is disabled process-wide.
- The server initializes the LLM provider at startup; model requests and the settings-page connectivity probe can send the key.
- The manifest declares no install, preinstall, or postinstall scripts.
- The README describes a local workstation and says users can configure an arbitrary OpenAI-compatible LLM endpoint.
- Provider initialization alone does not send a request; the key is transmitted when a model request or connectivity probe runs.
Source & flagged code
7 flaggedSource appears to send environment or credential material to an external endpoint.
server.jsView on unpkg · L6Findings
Affected versions and remediation
This report applies to aims-workstation@0.1.27.
The versions list below shows the verdict for each recorded version.
- Avoid installing aims-workstation@0.1.27. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.