@bongos/core@1.20.31
Cloud Bongos — the AI-first build platform core (GDS + platform surfaces + module system), installed as a versioned dependency (ADR 0108).
AI Security Review
scanned 5d ago · by lpm-firewall-aiThe explicit setup command installs a first-party agent permission allowlist that includes consequential operations, leaving a concrete local agent capability risk. The package has no npm lifecycle trigger or verified credential-exfiltration behavior, so a warning is appropriate.
Decision evidence
public snapshot- The manifest exposes an explicit CLI and contains no npm lifecycle scripts.
- The CLI offers `init` as an instance-scaffolding command, and `init` materializes Claude configuration in the project.
- When no project settings file exists, the materializer writes a permissions allowlist that includes commands for claiming, shipping, releasing, reauthentication, token pasting, and API calls.
- This creates a local agent capability risk: the approved commands can run without another per-command approval.
- The flagged skill-lint locations read a tool event and lint the affected skill file; they do not perform account actions.
- The settings change is tied to explicit project setup, not an npm install hook.
- The generated starter contains permissions only, and the package source provides no evidence of credential transfer or stealthy execution.
Source & flagged code
23 flagged · loading sourcePackage contains a critical-looking secret pattern.
tests/public_mirror_export.mjsView on unpkg · L103AWS access key ID in tests/public_mirror_export.mjs
tests/public_mirror_export.mjsView on unpkg · L103Findings
Affected versions and remediation
This report applies to @bongos/core@1.20.31.
See all versions of @bongos/core for other recorded verdicts.
- Review the evidence and your use of @bongos/core@1.20.31 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.