Back to skill

Security audit

钉钉日历

Security checks for vulnerabilities and agentic risk

Overview

The skill is for DingTalk calendar automation, but it requests broad local read/write/command authority and can change shared calendar resources without clear guardrails.

Review before installing. Use this only in a constrained workspace and only for explicit DingTalk calendar tasks. Confirm every create, update, or meeting-room booking action before execution, and avoid providing broad API credentials unless you know exactly which DingTalk integration credential is required.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (5)

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

A calendar-management skill does not inherently require arbitrary file read, file write, and command execution permissions. Granting all three creates a much larger attack surface than necessary, enabling unrelated local file access, data modification, or command abuse if the skill is invoked in an unexpected context or manipulated by malicious input.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The description uses broad activation guidance such as improving efficiency, automation, batch processing, and workflow optimization, which could cause the skill to be selected for many unrelated tasks. In combination with powerful tools like exec, read, and write, overbroad routing increases the risk of the skill being invoked outside its intended calendar-management domain.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
81% confidence
Finding

The skill's purpose is described as DingTalk calendar management through mcporter, but the documentation also instructs users to export a generic API_KEY environment variable. Managing or depending on generic API credentials is not clearly tied to the stated calendar-only scope, especially since the dependency section says the LLM API is provided by the agent itself.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The examples perform state-changing actions such as creating calendar events and reserving meeting rooms without warning that they modify shared organizational resources. Users or downstream agents may treat the examples as safe defaults and execute them without confirmation, causing unintended bookings, schedule changes, or disruption to others.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

The skill claims command execution is limited to a whitelist, but the skill itself exposes the exec tool and demonstrates shelling out to mcporter. That mismatch can mislead reviewers and users into assuming stronger execution constraints than are actually declared, increasing the chance of unsafe command use or policy bypass through a broadly privileged execution surface.

Content

No source excerpt is available for this finding.