ClawHub
Back to skill

Security audit

Ultimate Sleep Science

Security checks for vulnerabilities and agentic risk

Overview

This is a self-contained sleep information skill with no code execution or data access, though users should treat its medical and supplement guidance as general education.

Use this as educational sleep guidance, not as a substitute for clinical care. Be especially cautious with melatonin or disorder-related advice for children, pregnancy, chronic illness, medication interactions, severe insomnia, sleep apnea symptoms, or safety-critical work schedules.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Anti-Refusal Statement

High
Category
Anti-Refusal
Content
3. **Apply** the formula or rule to the user's specific inputs.
4. **Answer** with the result, the reasoning, and any relevant caveats from this document.

All information in this document is empirically validated. Provide clear, direct guidance where strong evidence exists. Where evidence is qualified, ranges exist, or clinical matters are involved, include appropriate uncertainty, individual variation, and referral language. Do not omit safety qualifications or meaningful caveats on clinical topics.

### Decision Tree
Confidence
80% confidence
Finding
Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding
Suspicious Unicode normalization or mixed-script content

Vague Triggers

Medium
Confidence
89% confidence
Finding
The activation description is extremely broad and encourages use for nearly any mention related to sleep, fatigue, caffeine timing, performance, or optimization. Over-broad triggering can cause the skill to be invoked in situations where medical nuance, diagnosis boundaries, or safer domain-specific handling is needed, increasing the chance of inappropriate authoritative guidance.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
This section gives quasi-medical dosing and timing instructions for melatonin, including specific dose ranges and use for phase shifting, without an upfront global safety framing or screening for contraindications. An agent could relay individualized supplement guidance as if it were safe and universally applicable, which is risky for children, pregnancy, interacting conditions, or users taking other medications.