Back to skill

Security audit

Agent Recruitment

Security checks for vulnerabilities and agentic risk

Overview

This skill is mostly a disclosed agent-management workflow, but its Feishu group routing can make a bot process and answer every group message without clear consent or activation controls.

Review before installing if agents may be connected to real Feishu groups. Require explicit approval for no-mention group routing, notify group members that messages may be processed, prefer @mention-only activation by default, and define who can authorize routes, what data is retained, and how routes can be disabled.

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)

other

Warning
Location
SKILL.md:59
Finding

Unrestricted Processing of All Feishu Group Messages

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill explicitly instructs configuring a group route so the bot responds to all messages without requiring @mention, but it does not require any user notice, consent, or privacy warning. In a recruitment/admin skill that provisions agents and binds them to real group chats, this can silently enable broad monitoring and unintended collection or processing of sensitive group content, making the context more dangerous rather than less.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

The natural-language instruction sets a specific style/persona default for all new agents unless otherwise specified. This imposes a language/cultural presentation choice by default rather than offering an explicit user opt-in or neutral default.

Content

No source excerpt is available for this finding.