ClawHub
Back to skill

Security audit

知识图谱基础版

Security checks for vulnerabilities and agentic risk

Overview

This knowledge-graph skill is mostly coherent, but it asks for broad command execution without clearly limiting when or how it should be used.

Review this skill before installing. It is not showing exfiltration or destructive intent, but it gives an agent command-execution capability for a task that mainly needs local graph reads and writes. Prefer installing only if you can restrict or monitor exec use, and store exported graph data only in intended local locations.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
94% confidence
Finding
The skill is described as a basic local knowledge-graph capability focused on CRUD, querying, and local persistence, yet the manifest grants exec in addition to read/write without any clear need or constraint. Unnecessary command execution materially expands the attack surface because an agent using this skill could be induced to run arbitrary system commands, leading to code execution or broader host compromise.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill explicitly supports persistent storage to local JSON/SQLite files and export/import operations, but it does not prominently warn users that data will be written to disk or discuss sensitivity, path selection, overwrite risk, or retention. In an agent context, silent local persistence can expose confidential graph data, create unintended artifacts on shared systems, or allow risky file-target choices if users are not clearly informed.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The manifest advertises both write and exec capabilities, but the skill description presents the feature set as ordinary graph management without corresponding safety caveats. In a skill consumed by an agent, undisclosed file-writing and command-execution permissions make the skill materially more dangerous because users and orchestrators may invoke it under the assumption that it only performs benign data operations.

Static analysis

No suspicious patterns detected.