ClawHub
Back to skill

Security audit

Using Creed

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed workflow-routing skill for software engineering tasks, with broad but non-destructive instructions and no code, persistence, credential use, or data access.

Install this only if you want Creed's engineering workflow to influence most software-development conversations in that workspace. It may slow lightweight tasks by requiring extra skill checks, but the inspected artifact does not show hidden code execution or sensitive data handling.

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 activation condition says to use this skill when starting any conversation in a Creed-enabled workspace, which is broad enough to seize control of most sessions regardless of the user’s actual task. In practice this can create instruction shadowing, where the skill inserts mandatory workflow and decision gates before ordinary work, increasing the chance of mis-prioritizing local policies or causing denial-of-service-like friction across unrelated conversations.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The 'even a 1% chance a Creed skill might apply' rule is an extremely low invocation threshold that strongly biases the agent toward always loading additional skills before acting. This is dangerous because it enables broad prompt capture and recursive policy expansion, making the agent defer to skill content in ambiguous cases and potentially override more relevant task-specific guidance with workflow mandates.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.