ClawHub
Back to skill

Security audit

reset-context-contamination

Security checks across malware telemetry and agentic risk

Overview

This is a prompt-only context-reset skill with no code execution or data access, though its broad activation wording could surprise users.

Install this if you want an agent to deliberately reset stale conversation framing. Be aware that ordinary phrases like asking for a fresh take or saying the discussion is going in circles may activate it, so users who want resets only by explicit command may prefer narrower trigger wording.

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 (1)

Vague Triggers

Medium
Confidence
92% confidence
Finding
The activation description is broad enough to match routine user phrases such as wanting a 'fresh take' or saying the conversation is 'going in circles,' which can trigger the skill outside of an explicit, intentional invocation. Because the skill tells the agent to discard prior framing and potentially recommend a new session or hand work to a fresh subagent, accidental activation can disrupt task continuity and bypass user expectations about how context should be used.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.