ClawHub
Back to skill

Security audit

Daily Tutor

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed daily-study helper that reads a local study list, records progress locally, and can optionally use an external quiz service if the user configures it.

Install only if you want a local file-backed study tracker. Do not put secrets, confidential notes, or proprietary material in data/data.json if you plan to generate Quizbuild quizzes, because those study items may be sent to the configured external MCP service.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (7)

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · README.md (reported line 8)May include surrounding context.

md
## Data Setup (`data.json`)

To use this skill, simply create (or update) the `data.json` file inside the `data/` directory with the content you want to learn. The system supports **any data structure**, so you are not restricted to any specific format!

1. Open the `data/data.json` file.
2. Write your learning content in a JSON Array format. In each element, you can freely name the data fields. For example:

External Transmission

Medium
Category
Data Exfiltration
Confidence
95% confidence
Finding

The skill instructs users to configure an external MCP server and send learning content to a third-party service for quiz generation. Because the skill supports arbitrary study data, users may unknowingly transmit sensitive or proprietary content outside the local environment, creating confidentiality and privacy risks.

Content

Scanner excerpt · README.md (reported line 51)May include surrounding context.

md
"mcp": {
  "servers": {
    "quizbuild": {
      "url": "https://api.quizbuild.com/mcp/quizbuild",
      "transport": "streamable-http",
      "headers": {
        "Authorization": "Bearer <TOKEN_API_QUIZBUILD>"

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The README says to "simply ask your agent" followed by a natural conversational phrase, "Let's review my daily items and give me a practice quiz." This is common everyday language and the file does not provide tighter trigger constraints or exclusion conditions, making unintended activation more plausible.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding

The skill invokes a local Python script that reads study data and updates progress tracking files, but the manifest does not declare any tool scope or permissions boundaries. This increases the chance of unintended file access or writes beyond what a user or orchestrator expects, especially because the skill description is generic and the script operates on local state.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill description is broad enough to match many ordinary education-related requests, which can cause the agent to invoke it in situations where the user did not intend local file-backed study tracking. Over-broad routing is dangerous because the skill reads and writes persistent user data and may disclose or modify local content under an innocuous-seeming prompt.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The 'When to Use' section contains permissive triggers like 'daily lesson' and cron-driven execution without clear constraints, which can lead to automatic activation and data mutation in contexts broader than intended. Because the skill tracks learned items and prevents repeats, accidental invocation has side effects on persistent progress state, not just harmless content retrieval.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

This JSON dataset presents meanings exclusively in Vietnamese (for example, repeated 'Chữ ...' entries) while teaching Japanese characters, with no indication that language choice is optional. The policy flags natural-language locale constraints when a skill or data file effectively enforces one language without offering user choice or documenting a justified regional scope.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.