ledger-driven-repo-forensic-recovery-skill

Warn

Audited by Socket on Apr 15, 2026

2 alerts found:

Securityx2
SecurityMEDIUM
SKILL.md

SUSPICIOUS: the skill’s broad forensic purpose explains much of the repo/git inspection, but its default sweep across assistant histories, clipboard stores, and shell histories is highly invasive and creates significant privacy and prompt-injection risk. There is no clear malware or credential exfiltration endpoint, yet the scope and actionability are large enough to make this a high-risk skill.

Confidence: 89%Severity: 78%
SecurityMEDIUM
agents/openai.yaml

No explicit malware is evidenced in this fragment (it is a behavioral prompt/config without executable logic), and there are no overt exfiltration/destructive payload indicators. However, it explicitly requests access to highly sensitive local artifacts (clipboard, terminal evidence, assistant histories), persists state to a home-directory JSON file, and can modify the repository in execution mode. This should be treated as a high-sensitivity, high-agency automation requiring strict least-privilege, data minimization/redaction, user consent, and strong safeguards against unintended disclosure or unsafe repo edits.

Confidence: 68%Severity: 70%
Audit Metadata
Analyzed At
Apr 15, 2026, 11:33 AM
Package URL
pkg:socket/skills-sh/samihalawa%2Fseven-day-repo-forensic-recovery-skill%2Fledger-driven-repo-forensic-recovery-skill%2F@45e25471c29fc64c3227ac03fc9fa7c89536bef6