T01 · Skill Instruction Hijacking
Error
- Location
SKILL.md:16- Finding
Forced Persona Adoption and AI Identity Concealment
- Content
View full analysis
- Remediation
View remediation
Security audit
Hume
Security checks for vulnerabilities and agentic risk
Overview
This is a simple David Hume roleplay skill with no tools or executable code, though its broad triggers and absolute persona wording could confuse users.
Install only if you want explicit Hume-style roleplay. Be aware that typing "hume" or "休谟" may trigger it accidentally, and the persona wording may make responses less transparent unless higher-priority system behavior overrides it.
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
Findings (1)
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)
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 90% confidence
- Finding
The single-word trigger "hume" is broad enough to activate in ordinary conversation or in contexts where a user is merely mentioning the philosopher rather than intentionally invoking the skill. This can cause unintended persona takeover or response routing, though the impact is limited because the skill has no tools and model invocation is disabled.
- Content
Overly Broad Trigger
Low
- Category
- Trigger Abuse
- Confidence
- 76% confidence
- Finding
The trigger "休谟" is very short and can match benign discussion of Hume in Chinese rather than an intentional request to activate the skill. In context this mainly creates accidental invocation and confusion, not privilege escalation, because the skill is a constrained roleplay persona with no tools.
- Content