ClawHub
Back to skill

Security audit

Refactor Agents Md

Security checks across malware telemetry and agentic risk

Overview

This skill is a documentation-only helper for reorganizing agent instruction files, with no hidden scripts or malware signals found.

Install this when you specifically want help auditing or restructuring AGENTS.md/CLAUDE.md guidance. Use review or proposal mode for sensitive repositories, and only use apply mode after you are ready for repository instruction files to be edited and reviewed in a diff.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
87% confidence
Finding
The skill description is broad enough to be invoked for generic cleanup, splitting, or review of AGENTS.md / CLAUDE.md guidance, which increases the chance of accidental selection in contexts where the user did not specifically intend repository-instruction refactoring. Because this skill can route into an in-repo rewrite workflow, over-broad matching can lead to unintended modification of governance or instruction files that affect future agent behavior.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill explicitly offers an 'apply the refactor in the repo' mode but does not prominently warn that this will modify repository files. In an agent environment, omission of a clear write warning can cause users or orchestrators to invoke a destructive workflow without realizing that AGENTS.md and related instruction files will be rewritten, potentially altering future agent constraints and project guidance.

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.