Back to skill

Security audit

brain-state-sleep-anxiety-cbti

Security checks for vulnerabilities and agentic risk

Overview

This is a text-only sleep and anxiety self-help skill with clear medical disclaimers and no hidden code, persistence, or data access.

Before installing, consider that this skill may activate for general sleep or anxiety conversations and will tend to answer in Chinese with CBTI-style self-help guidance. It should be treated as educational support only, not medical care, especially for severe, worsening, crisis, medication, or diagnostic concerns.

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
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 (2)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The description requires that the skill use plain-language explanations throughout and is written as a fixed Chinese-language interaction style, with no indication that users may choose another language. This can violate language or locale policy where skills should not force a specific language without user opt-in.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The trigger rules are extremely broad: any user asking about sleep or anxiety will automatically invoke this skill. In a health-adjacent context, this can cause the agent to override more appropriate general or crisis-sensitive handling and deliver prescriptive CBT-I style advice to users whose issues may be unrelated, severe, or require escalation. The danger is amplified because the skill is scoped to mental health and sleep concerns, where over-triggering can lead to misrouting, oversimplified guidance, or missed safety checks.

Content

No source excerpt is available for this finding.