Back to skill

Security audit

feishu-drive

Security checks for vulnerabilities and agentic risk

Overview

This Feishu Drive skill is a cloud-file management guide, but it also directs agents to make persistent sharing and public-edit permission changes without enough scoping or warnings.

Install only if you intend this skill to manage Feishu Drive permissions, not just files. Before use, require explicit user approval for every collaborator, full_access grant, external sharing change, and especially any internet-public read or edit link; prefer named collaborators and internal-only sharing by default.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The documentation explicitly includes internet-public readable and editable sharing modes, which can expose internal documents to anyone with the link or even broader audiences depending on platform behavior. In a cloud drive management skill, embedding public exposure options without strong justification, constraints, or warnings materially raises data leakage risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

Recommending one-click internet-public editing for vendor collaboration normalizes one of the riskiest sharing modes without requiring a safety review or warning. Public edit access can lead to unauthorized modification, malware staging, defacement, or broad disclosure of sensitive business materials.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The declared permissions describe basic file operations, but the skill text makes permission-management actions effectively mandatory after writes and discusses changing sharing visibility. This creates a capability and intent mismatch that can cause operators or downstream systems to underappreciate the real access-control effects of using the skill.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

Retrieving view/download records and managing security labels are specialized governance and compliance functions, distinct from the stated core purpose of uploading, downloading, moving, searching, and metadata management. These capabilities appear as extra administrative functions without corresponding purpose justification.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill is presented as routine file management, but the guidance expands it into access delegation and visibility modification, including broader sharing modes. That broader operational scope increases the chance of unintended exposure because users may invoke it expecting storage actions, not ACL or publication changes.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill describes internet-public permission settings without an explicit warning about confidentiality, unintended indexing, external resharing, or data-retention consequences. Users may apply high-exposure settings as routine workflow steps without recognizing that they are publishing content outside the organization.

Content

No source excerpt is available for this finding.