ClawHub
Back to skill

Security audit

goal-runner

Security checks across malware telemetry and agentic risk

Overview

This skill is an openly described roadmap automation workflow that can make code changes and optional git operations only within the stated user-directed task loop.

Install this only if you want an agent to work through a roadmap file end to end. Review goal text carefully: commits and pushes should be mentioned only when you really want them, and otherwise the skill is designed to leave changes for human git review.

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
91% confidence
Finding
The skill declares activation on phrases like "work on the roadmap tasks til completion," "work the roadmap until completion," and "launch sub agents to work the tasks," plus any /goal naming a roadmap/task file. Although domain-related, these triggers are still fairly broad and the file does not provide concrete negative examples beyond a couple of exclusions, which could lead to accidental invocation in adjacent task-management contexts.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
This markdown file describes an unattended loop that can rewrite task files, run tests, create branches, commit, and push when goal text authorizes it. While those behaviors are documented later in detail, there is no concise upfront warning near the description telling users that the skill may autonomously modify the working tree and perform git operations under certain conditions.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.