ClawHub
Back to skill

Security audit

Software Engineer for projects

Security checks across malware telemetry and agentic risk

Overview

The skill appears to be a project-coordination aid whose local file updates fit its purpose, but users should be aware it can modify workspace coordination documents.

Install if you want an agent to help maintain project state and coordination records. Before using it in a shared repo, make sure the team is comfortable with the skill updating local planning and audit files, and review diffs before committing those changes.

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
90% confidence
Finding
The trigger conditions are extremely broad, causing the skill to activate on generic engineering discussions such as code audit, architecture decisions, or branch review. In this skill's context, activation grants procedures for reading project state, checking queues, reviewing branches, and writing coordination files, so accidental invocation could lead to unintended file modifications or project-scoped actions without a clear user request.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to write and update multiple project files (STATE.md, KNOWN_ISSUES.md, RUNBOOK.md, queue files, implementation guides) but does not prominently warn the user that invoking the skill can modify workspace state. Because these files serve as coordination and audit records, an unintended invocation or prompt-injection-induced use of the skill could silently alter operational artifacts and mislead other agents.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.