jev-code-review

Warn

Audited by Socket on Sep 20, 2026

2 alerts found:

Securityx2
SecurityMEDIUM
SKILL.md

SUSPICIOUS: the skill’s stated review purpose is coherent, and it avoids silent installation, but it relies on a not-clearly-verifiable `jev-decide` CLI that receives an API key and submits review evidence to OpenRouter/providers. The main issue is dependency trust and credential forwarding, not overtly malicious behavior.

Confidence: 82%Severity: 83%
SecurityMEDIUM
assets/example.json

The diff definitely weakens the regression test by replacing its substantive parser-output assertion with `assert True`. The test receipt does not establish that the stated parser goal was achieved; it only establishes that the weakened test passed. This warrants direct review because the change conflicts with the acceptance condition. No executable malicious behavior is present in the supplied JSON.

Confidence: 99%Severity: 82%
Audit Metadata
Analyzed At
Sep 20, 2026, 05:01 AM
Package URL
pkg:socket/skills-sh/wuyoscar%2Fjev-skill%2Fjev-code-review%2F@f10dc27a376dd8c7eb3b7d62dfb6244bfeb73a909c914a860bff4f0e0c0d6a9d