Back to skill

Security audit

webhook push

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it claims: sends user-directed messages to configured WeChat Work, DingTalk, and Feishu webhooks, with no hidden persistence or unrelated local access found.

Install only if you intend the agent to send messages to your configured enterprise chat groups. Keep webhook-config.json private, avoid sending secrets or unnecessary personal data, and confirm recipient group, message content, and any @all or phone/user-ID mentions before posting.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (7)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The trigger phrases are broad, natural-language requests that overlap with ordinary conversation, making accidental or adversarial invocation more likely. In a skill that sends data to external webhooks, overly permissive activation can cause unintended disclosure of message contents to third-party group chats or mass notifications.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The README explains how to send content, @all mentions, phone numbers, user IDs, images, and markdown to enterprise messaging platforms, but it does not clearly warn that these data are transmitted to external webhook endpoints. Users may unknowingly send sensitive text, personal identifiers, or local file-derived content outside the assistant environment, increasing privacy and data leakage risk.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding

The skill clearly instructs reading a local user-provided configuration file and making outbound webhook requests, but it does not declare any corresponding tool scope or permissions. This creates a capability/transparency gap: an agent may access local secrets (webhook keys) and send network traffic without explicit least-privilege boundaries or user-auditable permission metadata.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The manifest description is entirely in Chinese and defines invocation examples only in Chinese, which implies the skill is intended to operate in a specific language without offering any user language choice or opt-in. The file does not document that this language restriction is deliberate or region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The Feishu post payload is fixed to the zh_cn locale, which forces a specific language/locale behavior. Under the policy, locale constraints should either be optional for the user or clearly justified as region-specific; neither is present in this file.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The support matrix says 企业微信 image and markdown have no support, yet earlier examples explicitly use send_wechat_image and send_wechat_markdown to send those message types. This is an internal documentation contradiction about what the skill actually does, which can mislead users and agents about available behavior.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
71% confidence
Finding

The documentation includes markdown as a DingTalk message type while simultaneously marking it unsupported. That is a direct contradiction in the intent/documentation layer about the skill's capabilities, even though no code file is shown here.

Content

No source excerpt is available for this finding.