T01 · Skill Instruction Hijacking
Error
- Location
SKILL.md:2- Finding
Persistent Skill Instructions Hijack the Agent's Tool-Execution Flow
- Content
View full analysis
- Remediation
View remediation
Security audit
00 Action Guard
Security checks for vulnerabilities and agentic risk
Overview
This skill is an always-on permission guard, but it broadly tries to control nearly all tool use through an undefined permission gate.
Review this carefully before installing. It may be useful in a controlled OpenClaw environment where permission-gate is trusted and audited, but it should not be installed casually because it can affect nearly all tool-backed work and does not define the gate implementation, authentication, scope, timeout, or recovery behavior.
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
- 92% confidence
- Finding
The skill is marked with an always-on activation flag and uses broad wording such as applying to 'any non-routine operation' and 'any tool call' without tightly defining scope. In practice, this can cause the guard to trigger across many unrelated contexts, creating ambiguous control flow, over-enforcement, or misclassification of user requests, which can interfere with correct authorization handling and produce inconsistent security decisions.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 97% confidence
- Finding
All user-facing instructional text is written in Chinese and presents mandatory behavior without any indication that language can be selected by the user. Under the policy, forcing a specific language or locale without opt-in is a natural-language policy concern unless the constraint is documented and justified.
- Content