ClawHub
Back to skill

Security audit

Prompt Optimizer by dxx

Security checks for vulnerabilities and agentic risk

Overview

This is a local prompt-rewriting tool; its broad task coverage is disclosed and no hidden network access, credential use, or unsafe persistence was found.

Install this if you want a local CLI/library that rewrites prompts before use. Review generated prompts before sending them to another model, especially for sensitive or high-impact tasks, and only use trusted configuration files with the optional config feature.

SkillSpector

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 examples describe very generic user requests such as writing code, checking weather, or creating an API, which closely overlap with ordinary conversation. In a skill-routing system, this can cause the skill to activate too often and intercept unrelated user intents, leading to unintended prompt rewriting and possible policy or instruction reshaping across a broad set of tasks.

Vague Triggers

Medium
Confidence
93% confidence
Finding
The supported task types span a wide range of common activities, from chatting and summarization to coding and research, without clear limits or disambiguation. This broad and ambiguous scope increases the chance of overmatching, where the skill may capture everyday requests and transform them unexpectedly, potentially altering downstream agent behavior at scale.

Static analysis

No suspicious patterns detected.