@ngockhoale/ukit@3.4.43
Install/update an index-first AI workspace for Claude Code, OpenAI Codex and omp (Oh My Pi).
AI Security Review
scanned 10d ago · by lpm-firewall-aiThe source shows first-party agent setup that runs through an explicit CLI install command, so it merits a warning under the agent-extension policy. I found no source evidence supporting a publish block.
Decision evidence
public snapshot- The package exposes a CLI that routes an explicit `ukit install` command to its installer; the manifest has no install lifecycle hook.
- The installer applies package-owned agent templates to project configuration. A reviewer agent is enabled by default and installed with backup-on-overwrite behavior.
- The flagged benchmark and probe targets perform fixture, measurement, or workspace-inventory work; the installer targets include protected writes, root-level ignore rules, guarded gateway defaults, hook repair, and bounded uninstall paths.
- Gateway resilience settings are applied only after a custom gateway is detected, and the source says existing non-default user values are preserved.
- No npm install hook automatically performs the agent setup; it is triggered by running the CLI installer.
- The inspected source does not establish credential theft, remote code execution, or unconsented broad agent-control takeover.
Source & flagged code
17 flagged · loading sourcePackage source references dynamic require/import behavior.
scripts/bench/runGold.mjsView on unpkg · L121Source 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@3.4.43.
See all versions of @ngockhoale/ukit for other recorded verdicts.
- Review the evidence and your use of @ngockhoale/ukit@3.4.43 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.