Back to skill

Security audit

Maxhub Zhihu

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious with medium confidence, mainly because its Zhihu workflow includes fallback instructions for Douyin APIs, alongside an unpinned installation command.

Review the Douyin fallback and broad activation examples before installing, and check that the installation command uses a version you trust. The scan did not establish malicious intent, but its analysis is marked partial because two references were unresolved, so the results do not establish that every referenced behavior was assessed.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • 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
Findings (6)

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

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding

The installation command invokes an unpinned npm CLI, allowing the code executed during installation to change over time. A compromised future package release could execute with the installing user's permissions; this is a supply-chain exposure, not evidence that the package is malicious or an MCP server.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding

The documented npx command does not pin the clawhub package version, so installation can execute a different release without review. If a selected release is compromised, its code could access local files and credentials with the installer's permissions; the command itself does not establish malicious intent.

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/zhihu/{endpoint}?%7Bparams=}" \
  -H "$maxhub_auth_header"

# POST example

The fallback action table directs calls to Douyin APIs despite the manifest limiting the skill to Zhihu data queries.

Medium
Category
Excessive Agency
Confidence
92% confidence
Finding

The executable fallback guidance authorizes Douyin endpoints inside a Zhihu-only workflow, creating a scope-control defect that could produce unrelated authenticated requests, disclose query parameters outside the intended service scope, or consume API quota. The mappings are keyed to Douyin endpoints rather than normal Zhihu endpoints, which limits their direct reachability, but their inclusion as authoritative fallback instructions can still misdirect an agent.

Content

No source excerpt is available for this finding.

Example invocation prompts use generic terms without specifying Zhihu.

Low
Category
Trigger Abuse
Confidence
93% confidence
Finding

The example prompts include standalone terms such as 用户 (user), 资料 (information), 搜索 (search), and 文章 (article), which overlap with ordinary requests unrelated to Zhihu. Although the README identifies the skill as a Zhihu assistant, it does not constrain these examples to Zhihu-related requests or give exclusion conditions.

Content

No source excerpt is available for this finding.

The example invocation phrases are generic words that could trigger the skill outside Zhihu-related requests.

Low
Category
Trigger Abuse
Confidence
93% confidence
Finding

The usage examples present standalone words such as “用户” (user), “资料” (information), “搜索” (search), and “文章” (article) as commands. Although the introduction identifies this as a Zhihu assistant, the invocation examples do not require Zhihu context or give exclusions, leaving overlap with ordinary requests and other skills.

Content

No source excerpt is available for this finding.