@ngockhoale/ukit@3.4.31
Install/update an index-first AI workspace for Claude Code, OpenAI Codex and omp (Oh My Pi).
AI Security Review
scanned 12d ago · by lpm-firewall-aiStatic inspection found conditional credential exposure: the decision URL is independently configurable, while key resolution can fall back to the shared Anthropic token, and no destination allowlist is enforced. The decision stage gates use of this client, and the default endpoint is package-aligned, so the finding warrants a warning rather than a block.
Decision evidence
public snapshot- src/core/gatewayProbe.js accepts a configurable decision URL and can fall back to a shared UNIC token; src/decision/client.js sends the resolved authorization header to that URL.
- The decision client is reachable when the VM decision stage is enabled; package.json exposes a CLI and declares no install lifecycle hook.
- The required agent-related files show benchmark and probe code, guarded workspace maintenance, hook repair, and a watchdog delegating to its coordinator; no separate concrete malicious behavior was established.
- The decision request uses the package's intended UNIC decision service by default, and runtime calls are stage-gated; the exposure requires a configured alternate endpoint.
Source & flagged code
6 flagged · loading sourcePackage source references dynamic require/import behavior.
scripts/bench/runGold.mjsView on unpkg · L121Package source references weak cryptographic algorithms.
Findings
Affected versions and remediation
This report applies to @ngockhoale/ukit@3.4.31.
See all versions of @ngockhoale/ukit for other recorded verdicts.
- Review the evidence and your use of @ngockhoale/ukit@3.4.31 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.