@bongos/core@1.20.16
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-aiSource inspection confirms transcript content reaches the configured Anthropic CLI after a setup prompt that says raw transcripts remain local. Consent and a time window limit activation, but the prompt's disclosure leaves a real unresolved privacy risk, so this warrants a warning rather than a publish block.
Decision evidence
public snapshot- Interactive setup says only a distilled summary is shared and transcripts stay local, but treats an empty answer as yes.
- Saved consent gates transcript staging and analysis; the analyzer reads transcripts and places a compacted transcript in its prompt.
- The prompt is passed to the configured Anthropic CLI provider; afterward, derived analysis is sent to the GDS inbox.
- The manifest has no install lifecycle scripts, and the other account-action hints resolve to metrics and console output.
- The agent-related paths concern project setup, local configuration, worktree naming, and module scaffolding.
- Analysis requires a saved consent decision within a time window; transcripts are first staged locally.
- The GDS inbox receives the derived analysis payload, not the raw transcript directly.
- There is no package install hook in the inspected manifest.
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@1.20.16.
See all versions of @bongos/core for other recorded verdicts.
- Review the evidence and your use of @bongos/core@1.20.16 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.