@ngockhoale/ukit@3.4.49
Install/update an index-first AI workspace for Claude Code, OpenAI Codex and omp (Oh My Pi).
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe source shows a default-enabled runtime path that forwards a host authentication token to a configurable gateway, with no destination allowlist or recipient-specific consent gate. The intended-provider description and sensitive-state filtering reduce certainty about misuse, but do not prevent exposure of the authentication token.
Decision evidence
public snapshot- The manifest exposes a CLI and defines no install lifecycle hook.
- The decision plane is enabled by default and task routing can invoke it during ordinary routing.
- Without a decision-specific key, the client falls back to a host UNIC token and sends it as authentication to the resolved gateway; the default gateway is openai.unicjsc.com, and the endpoint can be overridden without a destination allowlist.
- The request filters sensitive state content, but that filter does not gate the separate authentication header.
- The inspected agent integration locations concern project configuration, indexing, cleanup, or probes; they do not establish a separate agent-control attack.
- The source describes the gateway as the intended UNIC decision provider, and the decision feature can be disabled or configured with a different endpoint.
- The package has no install lifecycle hook, and the inspected state-packet path blocks content marked sensitive before transmission.
Source & flagged code
11 flagged · loading sourcePackage source references dynamic require/import behavior.
src/core/runtimeConfig.jsView on unpkg · L9Source file is highly similar to a previously finalized malicious package; route for source-aware review.
scripts/release/verify-release.mjsView on unpkgFindings
Affected versions and remediation
This report applies to @ngockhoale/ukit@3.4.49.
See all versions of @ngockhoale/ukit for other recorded verdicts.
- Review the evidence and your use of @ngockhoale/ukit@3.4.49 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.