ClawHub
Back to skill

Security audit

memory-latch

Security checks across malware telemetry and agentic risk

Overview

This instruction-only continuity-recovery skill is benign, with disclosed local persistence and no executable code or default credential/network use.

Install this if you want a local continuity log and stricter confirmation around irreversible actions. Review or replace the bundled manifest before relying on it, keep .memory-latch files free of secrets and sensitive personal data, and leave optional HMAC or wallet modes disabled unless you intentionally configured them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The trigger phrase "start over" is broad and commonly appears in normal conversation, so the skill may enter recovery mode when the user did not intend a continuity-reset workflow. In this skill, unintended invocation can degrade reliability and potentially gate or delay legitimate actions by forcing reconstruction or consent flows at the wrong time.

Vague Triggers

Medium
Confidence
85% confidence
Finding
The runtime trigger criteria include vague conditions like "missing expected state" and general tool or cache issues, which can be satisfied in many benign situations. Because this skill changes behavior around trust, recovery, and potentially irreversible-action gating, ambiguous activation can cause unnecessary fail-closed behavior, workflow interruption, or user confusion.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal