Back to skill

Security audit

Maxhub Weibo

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because it reports a mismatch between its stated Weibo query purpose and observed behavior, along with instructions that may route requests outside that scope; the scan is partial.

Consider the suspicious verdict and Skillspector's DO_NOT_INSTALL recommendation before installing. The scan covered all listed components but remains partial because two local references could not be resolved; the separate AIG scan completed with no findings, which does not resolve the concerns or establish safety.

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 Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied chunk is solely an environment-variable presence check, not the declared Weibo data-query functionality. Checking an API key could be a supporting implementation detail, and the code neither reveals the key nor transmits data. Nevertheless, the supplied code contains no implementation of the declared primary purpose, making its observed behavior materially different from the description.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx clawhub'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The installation command invokes an unpinned npm package through npx, potentially downloading and executing a changing release with the user's privileges. Although this is an installer rather than an MCP server reference, a compromised upstream release could expose local files and credentials.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx clawhub'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The installation command invokes an unpinned npm package through npx, allowing the executed installer to change as upstream releases change. This is not an MCP server configuration, but a compromised installer release could execute code with the user's privileges and access local secrets.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 53)May include surrounding context.

md
maxhub_auth_header="Authorization: Bearer ${MAXHUB_API_KEY}"

# GET example
curl -s "https://www.aconfig.cn/api/v1/weibo/{endpoint}?%7Bparams=}" \
  -H "$maxhub_auth_header"

# POST example

The fallback table directs requests to Douyin APIs despite the skill being scoped to Weibo data queries.

Medium
Category
Excessive Agency
Confidence
93% confidence
Finding

The fallback instructions explicitly endorse Douyin endpoints in a Weibo-only workflow, creating an unsafe cross-platform routing path if the agent applies the table. Combined with autonomous execution, this could send query parameters to an unintended API product, consume paid calls, or produce misleading results; the mappings' non-Weibo source endpoint names make actual triggering uncertain.

Content

No source excerpt is available for this finding.

Example invocation prompts include generic terms without a Weibo-specific scope requirement.

Low
Category
Trigger Abuse
Confidence
88% confidence
Finding

Examples such as “评论” (comments), “关注” (follow), “频道” (channel), and “推荐” (recommendations) are common outside Weibo. Although the README describes a Weibo assistant, it does not require Weibo context for these short prompts or provide exclusions, leaving their activation scope ambiguous.

Content

No source excerpt is available for this finding.

Usage examples include generic invocation phrases without a clear Weibo-specific activation boundary.

Low
Category
Trigger Abuse
Confidence
86% confidence
Finding

The example commands include standalone terms such as “评论” (comments), “关注” (follow), and “推荐” (recommendations), which also occur in ordinary requests unrelated to Weibo. Although the README describes a Weibo assistant, it does not explicitly restrict these generic examples to an established Weibo context or provide exclusion conditions.

Content

No source excerpt is available for this finding.