ClawHub
Back to skill

Security audit

Self Reflection

Security checks across malware telemetry and agentic risk

Overview

This skill is a documented local self-reflection helper that stores user-directed lessons on disk, with no hidden code or network behavior in the artifact.

Before installing, decide where the memory file should live and do not log secrets, customer data, credentials, or sensitive internal details. If you follow the README's external GitHub clone and CLI symlink steps, review that repository's executable code separately because this artifact itself does not include the CLI implementation.

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 (3)

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The markdown explicitly says the skill will 'track mistakes, log lessons learned, and build institutional memory over time,' and later documents auto-created memory/state files. This behavior affects user or agent data retention, but the README does not warn that entries may contain sensitive operational details and will be persisted on disk.

Vague Triggers

Medium
Confidence
86% confidence
Finding
The description says the agent performs reflection through "regular heartbeat-triggered reflections," but it does not clearly define when the skill should or should not activate beyond a general heartbeat concept. This ambiguity is reinforced by the required HEARTBEAT.md instruction later, which could cause the skill to run broadly whenever a heartbeat occurs without negative examples or scope limits.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The markdown instructs users to configure persistent storage files for reflections and state, including a memory file that will accumulate logged lessons over time. The documentation does not warn users that operational mistakes, lessons, or other potentially sensitive content will be stored on disk, which has privacy and data-retention implications.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.