ClawHub
Back to skill

Security audit

PR Advocacy

Security checks across malware telemetry and agentic risk

Overview

This PR automation skill is not clearly malicious, but it gives an agent broad authority to monitor PRs, change branches, and persist tracking data without clear user approval boundaries.

Install only if you want an agent to actively manage PR review work. Before use, restrict it to specific repositories and PRs, require confirmation before commits, pushes, comments, PR edits, or tracking-file writes, and replace or disable the hard-coded local memory path unless you explicitly want PR metadata stored there.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill explicitly instructs the agent to automatically commit changes to a PR branch in response to reviewer feedback, which modifies repository state without any explicit user confirmation or warning. In an autonomous agent context, this can cause unintended code changes, branch mutations, or policy violations, especially if reviewer comments are misinterpreted or adversarial.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill specifies a concrete local file path and directs the agent to persist PR tracking data to disk in real time, but does not clearly warn that it performs persistent local writes. This creates risk of silent filesystem modification, unintended data retention, leakage of sensitive repository metadata, and corruption or overwrite of user-local state if the path is shared or mis-scoped.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.