Skills
skills/stevengonsalvez/agents-in-a-box/coding-agent/Socket

coding-agent

Warn

Audited by Socket on May 5, 2026

1 alert found:

Security
SecurityMEDIUM
SKILL.md

SUSPICIOUS: the skill’s purpose is coherent, but it gives local coding agents broad authority over repositories and GitHub actions, including dangerous no-approval modes and autonomous push/comment flows. Main risks are agent autonomy, exposure of repo/PR data to third-party CLIs, and prompt-injection hazards when reviewing untrusted PR content.

Confidence: 86%Severity: 74%
Audit Metadata
Analyzed At
May 5, 2026, 09:08 AM
Package URL
pkg:socket/skills-sh/stevengonsalvez%2Fagents-in-a-box%2Fcoding-agent%2F@983238d34501a75f068ea5457e7ff50612896f85