ClawHub
Back to skill

Security audit

Personality Distiller

Security checks across malware telemetry and agentic risk

Overview

This markdown-only skill has a coherent persona-building purpose, but it can read sensitive personal files and overwrite persistent agent persona files with too little scoping and rollback guidance.

Install only if you intentionally want an agent to generate or replace workspace persona files. Back up the six existing persona files first, give it only specific files or folders you are comfortable having analyzed, avoid secrets or highly sensitive personal material unless deliberate, and require an explicit confirmation before any overwrite.

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)

Vague Triggers

Medium
Confidence
94% confidence
Finding
The trigger scenarios include broad, everyday phrases such as wanting to improve decision quality or needing a thinking advisor, which can cause the skill to activate for ordinary conversation rather than an explicitly consented persona-distillation task. In this skill's context, that broad activation is more dangerous because execution can lead to deep web research, local file analysis, and overwriting workspace persona files, making accidental invocation materially risky.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill explicitly supports analyzing local corpus files to infer themes, worries, aspirations, and blind spots, but it does not provide a comparably clear warning about sensitive local data access as it does for overwriting persona files. This creates a privacy and consent risk because users may not understand that diaries, notes, or other sensitive workspace content could be read and distilled into a persona artifact.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The framework-selection guide uses very broad keyword triggers such as 'thinking', 'decision', 'system', and 'distill/extract XX', which can match ordinary user language and cause the skill to over-apply or misapply frameworks. In a skill that automatically distills and overwrites persona files, this weak routing logic increases the chance of generating inappropriate personas, confusing behavior, or steering outputs in unintended directions based on incidental wording rather than actual user intent.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.