canary

Warn

Audited by Socket on Sep 30, 2026

1 alert found:

Security
SecurityMEDIUM
SKILL.md

SUSPICIOUS. The core monitoring behavior matches the stated purpose, and data largely stays local or goes only to the user-specified site. The main issue is install/execution trust: the skill relies on numerous repo-local scripts plus a locally built browser binary from a personal GitHub repo, which is a disproportionate trust anchor for a monitoring skill even though it is same-project consistent. No clear credential harvesting, covert exfiltration, or malicious pre-execution behavior is present in the provided text.

Confidence: 86%Severity: 74%
Audit Metadata
Analyzed At
Sep 30, 2026, 07:02 AM
Package URL
pkg:socket/skills-sh/garrytan%2Fgstack%2Fcanary%2F@e263a128d32e531d5bdf802541b168a266dd6d24b8a48a77b408eab09a896227
Security Audit — socket — canary