Security audit

微博热搜

Security checks for vulnerabilities and agentic risk

Overview

This skill coherently fetches public Weibo trending topics through RedFox, but users should notice that it can use an API key and optionally create a recurring daily push task.

Install only if you are comfortable providing a RedFox API key and allowing calls to redfox.hk. If you enable the daily push feature, confirm the schedule intentionally and make sure you know how to disable the scheduled task in your OpenClaw environment.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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

Findings (11)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 93% confidence
Finding: The skill instructs use of an API key and execution of a Python script that necessarily consumes environment variables and performs outbound network access, but it does not declare those capabilities. Undeclared env/network use weakens user and platform visibility into what the skill can access and transmit, increasing the risk of unnoticed credential use or unexpected data egress.

Description-Behavior Mismatch

Medium

Confidence: 91% confidence
Finding: The README advertises a scheduled daily subscription/push capability even though the skill metadata describes only on-demand querying and analysis of current Weibo hot searches. This creates a scope mismatch that can mislead users into authorizing or expecting background behavior the skill may not be permitted to perform, increasing the risk of unintended task creation or privilege misuse.

Description-Behavior Mismatch

Medium

Confidence: 92% confidence
Finding: The README advertises a daily subscription/push feature that is outside the stated hot-search lookup purpose, which can mislead users and reviewers about the skill’s actual capabilities. In agent ecosystems, undocumented scheduling or subscription behavior is security-relevant because it may cause persistent actions, unexpected notifications, or future data access beyond the user’s immediate request.

Description-Behavior Mismatch

Medium

Confidence: 89% confidence
Finding: The output example normalizes automatic daily push behavior as if it were a standard feature, despite that behavior not being described in the stated skill scope. This kind of documentation mismatch is dangerous because users may be induced to authorize or expect persistent background actions without a clear security review of how subscriptions are created, stored, and revoked.

Context-Inappropriate Capability

Medium

Confidence: 91% confidence
Finding: The skill expands from one-time data retrieval into creation of a recurring scheduled task that repeatedly runs and pushes results. Persistent automation changes the trust boundary: a user may authorize what seems like a simple lookup but end up with ongoing execution, notifications, and repeated external API use, which can incur privacy, spam, or cost consequences.

Vague Triggers

Medium

Confidence: 83% confidence
Finding: Telling users to 'just describe your need in plain language' makes invocation boundaries ambiguous and can cause the skill to activate on loosely related conversational text. In an agent environment, overly broad activation guidance raises the chance of unintended execution, especially when the skill can make external API calls or perform analysis automatically.

Vague Triggers

Medium

Confidence: 87% confidence
Finding: Example trigger phrases such as 'Weibo trending' and 'What's hot today' are generic and overlap with ordinary conversation, making accidental activation more likely. Because this skill reaches out to an external service and returns live data, unintended triggering can cause unnecessary data access, confusing behavior, or action chaining in a broader agent workflow.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: Additional examples like 'What's trending on Weibo today' and 'Show me what's trending' remain broad enough to match normal user speech rather than deliberate skill invocation. In context, this skill is low-to-moderate sensitivity, but ambiguous triggers still increase the risk of unintended external calls and inconsistent agent behavior.

Vague Triggers

Medium

Confidence: 78% confidence
Finding: The trigger phrases are broad natural-language examples such as generic hot-search requests and a bare confirmation phrase, which can increase the chance of unintended activation or ambiguous routing. This is especially risky when combined with a subscription flow, because a vague phrase like '确认订阅' may be matched out of context and lead to persistent behavior the user did not clearly intend.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The README mentions creating a daily scheduled subscription but does not clearly warn users that this is persistent behavior continuing beyond the current session. Lack of a prominent warning undermines informed consent and can result in surprise notifications, unwanted recurring actions, or silent retention of user preferences.

Missing User Warnings

Low

Confidence: 84% confidence
Finding: The documentation asks users to place a live API key in config or environment variables without warning about secret handling, least privilege, rotation, or leakage through logs/shell history. While common, this can lead to accidental credential disclosure, especially in shared environments or when users copy configs into insecure locations.

Static analysis

No suspicious patterns detected.