ClawHub
Back to skill

Security audit

Skill Manager

Security checks for vulnerabilities and agentic risk

Overview

This skill clearly does what it claims, but it can persistently change which skills multiple agents load and includes under-scoped automatic sync and migration behavior.

Install only if you want one skill to manage skill availability across multiple agents. Before running it, inspect ~/.skills/agents.yaml and each skill's meta.yaml, use --dry-run for sync and migration, avoid universal scope for unreviewed skills, and back up agent skill directories before migration.

SkillSpector

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 (3)

Vague Triggers

Medium
Confidence
83% confidence
Finding
The handoff trigger is described broadly as 'when a new skill is detected or when you are handed off,' without clear authorization, validation, or user-confirmation boundaries. Ambiguous invocation conditions can cause the agent to act on unintended skills or contexts, increasing the chance of unauthorized provisioning steps being performed automatically.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The instructions direct the agent to run a synchronization script that modifies the filesystem by creating symlinks, but they do not require warning the user, previewing changes, or obtaining confirmation first. In an agent setting, filesystem-modifying actions without explicit consent or safety checks can lead to unintended persistent changes across agent environments.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The script performs a destructive `rm -rf` on an agent skill directory during migration without any explicit confirmation or safety interlock. Although the path is quoted and derived from enumerated directories, a migration mistake, unexpected directory contents, or stale state can cause irreversible loss of local skill data.

Static analysis

No suspicious patterns detected.