ClawHub
Back to skill

Security audit

Package Skill

Security checks for vulnerabilities and agentic risk

Overview

This skill is meant to reorganize other skills, but it can delete original skill folders without confirmation, so users should review it before installing.

Install only if you are comfortable with a tool that reorganizes and removes top-level skill folders. Back up the skills directory first, run commands manually with exact skill names, and verify copied package contents before relying on the packaged version.

SkillSpector

Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding
The skill documentation clearly instructs use of scripts that create, copy, move, remove, and update files and directories, but it does not declare corresponding permissions. That mismatch weakens trust boundaries and review controls because an agent may perform filesystem actions without an explicit permission contract, especially in a packaging tool that rewrites the skill library structure.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The workflow states that package creation will remove the original top-level skill directories, but it does not present this as a prominent warning, confirmation step, or backup requirement. In this context the skill is specifically designed to reorganize many skills at once, so silent destructive behavior increases the chance of accidental data loss, broken references, or unintended removal of important skills.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The create command copies the selected skills into the new package and then unconditionally deletes the original top-level skill directories with shutil.rmtree(). This is dangerous because a user mistake, stale path selection, or partial package-generation issue can cause irreversible data loss without any confirmation, dry-run, or rollback path.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The add command copies a skill into sub/ and then immediately removes the original directory with shutil.rmtree() without asking the user to confirm. This can lead to accidental destruction of a skill if the operator misunderstood the command semantics or if later update steps fail, leaving the environment in an inconsistent state.

Static analysis

No suspicious patterns detected.