ClawHub
Back to skill

Security audit

quality-gate

Security checks across malware telemetry and agentic risk

Overview

This skill is coherent as an egregore quality gate, but it can automatically commit code and post GitHub PR reviews from broad triggers like “quality” and “review.”

Install this only if you want an egregore-specific quality gate that may change and commit code and may post GitHub PR approvals, comments, or change requests. Prefer explicit invocation, verify the target branch and PR number, and consider narrowing the triggers or adding confirmation/dry-run behavior before using it in shared repositories.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Vague Triggers

Medium
Confidence
94% confidence
Finding
The trigger phrase "review" is broad and likely to match routine user requests that are not intended to invoke this skill. In this skill's context, accidental activation is more dangerous because the workflow can run repository checks, invoke other skills, and in PR-review mode post GitHub reviews/comments, causing unintended side effects.

Vague Triggers

Medium
Confidence
91% confidence
Finding
The trigger phrase "quality" is ambiguous and can overlap with many normal conversations about code quality or product quality. Because this skill orchestrates multi-step actions and may modify branches or interact with GitHub, an overly generic trigger increases the chance of unintended execution.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill describes posting GitHub reviews and inline comments in PR-review mode but does not present this as an explicit user-facing warning or consent checkpoint. That omission is risky because users may invoke the skill expecting local analysis only, while it can perform externally visible actions that affect collaboration workflows and repository history.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The self-review workflow includes an auto-fix loop that attempts fixes and commits them to the branch, but this behavior is not clearly surfaced as a warning. Automatic code changes and commits are significant side effects; if users do not expect them, the skill may alter source history or introduce changes without informed approval.

Shadow Command Trigger

Medium
Category
Trigger Abuse
Confidence
95% confidence
Finding
Using "review" as a trigger conflicts with a built-in command of the same name, creating a shadowing risk where the skill may intercept requests meant for the trusted built-in flow. In this case the danger is elevated because the skill can orchestrate additional tools, make repository changes, and post PR reviews, so command confusion can lead to unintended actions.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.