@melaya/runner@1.0.83
Run Melaya AI pipelines locally with your own LM Studio or Ollama models
AI Security Review
scanned 4mo ago · by lpm-firewall-aiSource inspection confirms intentional server-directed code execution rather than a concrete malicious install-time chain. Flag as suspicious because the remote-control capability runs with local user privileges and reads third-party CLI authentication state.
Decision evidence
public snapshot- `dist/connection.js` accepts `runner:run`, writes server-supplied Python, and spawns it.
- `dist/sharedVendor.js` downloads and writes a runtime Python bundle under `~/.melaya-runner`.
- `dist/detect.js` reads `~/.claude/.credentials.json` and `~/.codex/auth.json`.
- `dist/lumaLogin.js` invokes `npx playwright install chromium` at runtime.
- `package.json` has no preinstall/install/postinstall hook; only `prepublishOnly`.
- Remote bundle and pipeline content require HMAC verification using the supplied runner token.
- Credential reads are used for local model detection; no source path sends their raw contents to Melaya.
- Writes are limited to runner cache and temporary run directories; no AI-agent config mutation found.
Source & flagged code
5 flaggedPackage source invokes a package manager install command at runtime.
dist/lumaLogin.jsFindings
Affected versions and remediation
This report applies to @melaya/runner@1.0.83.
See all versions of @melaya/runner for other recorded verdicts.
- Review the evidence and your use of @melaya/runner@1.0.83 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.