skills/catena-labs/skills/panel-review/Socket

panel-review

Warn

Audited by Socket on May 11, 2026

1 alert found:

Security

SecuritySKILL.md

MEDIUM

SecurityMEDIUM

SKILL.md

The skill intentionally feeds untrusted PR diffs and review comments to external agents, and committed-target mode grants those agents write/exec access in isolated worktrees. That is a real indirect prompt-injection surface, though it fits the review purpose and is not covert exfiltration.

Confidence: 88%Severity: 72%

Audit Metadata

Analyzed At

May 11, 2026, 09:22 PM

Package URL

pkg:socket/skills-sh/catena-labs%2Fskills%2Fpanel-review%2F@31238e0d9cccc452be997975191509b1d38b73b6