T01 · Skill Instruction Hijacking
Warning
- Location
SKILL.md:23- Finding
Untrusted Group Messages Are Embedded Directly into Sub-Agent Instructions
- Content
View full analysis
- Remediation
View remediation
Security audit
feishu-routing
Security checks for vulnerabilities and agentic risk
Overview
This skill is a Feishu chat router, but it automatically sends group messages to sub-agents and forwards replies without clear safeguards against malicious or sensitive chat content.
Review this before installing in any real Feishu workspace. It should be limited to intended groups and least-privilege sub-agents, and the router should treat group messages as untrusted data, validate senders, redact sensitive content where possible, and require review before forwarding sensitive or high-impact responses. I found no evidence of malware, persistence, credential theft, or destructive 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
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 92% confidence
- Finding
The skill explicitly forwards full group chat message content and group context to sub-agents via
sessions_spawn, which creates unnecessary data sharing across agents without any notice, consent boundary, or minimization rule. In a chat-routing context this is materially risky because group messages may contain sensitive internal discussions, and every routed message expands exposure to additional agents and their logs.- Content
Natural-Language Policy Violations
Low
- Category
- Not specified by scanner
- Confidence
- 81% confidence
- Finding
All user-facing instructions and labels in this skill are presented exclusively in Chinese, and the document does not indicate that the language is optional or limited to a justified region-specific deployment. Under the language/locale policy, forcing a single language without user opt-in can be a policy violation.
- Content