Back to skill

Security audit

Tiktok数据助手

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious with medium confidence because scanners report undisclosed local legal-confirmation records and guidance that can trigger upgrades or requests beyond the stated API boundary.

Consider avoiding installation until the package and skill versions are pinned and reviewed, the upgrade guidance requires your explicit approval, CDN access is constrained to trusted destinations without forwarding credentials, and the legal-confirmation records no longer assert a review or consent that was not verified. Skillspector reports partial analysis because some local references were missing or ambiguous, so the evidence is incomplete; the other scanner completed with no findings.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
Findings (11)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied chunk performs no TikTok queries or API requests. Instead, it records a confirmed legal-compliance state. Local bookkeeping alone could be a supporting implementation detail, but asserting first-time legal confirmation without a confirmation step is a substantive behavior not disclosed by the description. This assessment is limited to the supplied chunk.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
93% confidence
Finding

The mismatch is not merely the use of a local log as an implementation detail. The code records a completed legal-review event, an undeclared compliance-state change unrelated to the described TikTok queries. The append succeeds only if the destination directory exists and is writable. This assessment is limited to the supplied chunk.

Content

No source excerpt is available for this finding.

YARA rule 'exploit_framework': Exploit framework components and payloads [hacktools]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · references/atoms/_index.md (reported line 10)May include surrounding context.

md
��应主动调用**。
> 仅在用户明确授权 + 一次性确认后才能执行。具体策略见 SKILL.md 的「高风险能力清单」章节。

| atom_id | 风险类别 | 限制原因 |
|---------|---------|---------|
| 🚨 extract_aweme_id_batch | bulk_extraction | bulk identifier extractor |
| 🚨 open_app_message | private_messaging | triggers private-message UI; abuse vector for spam/social-engineering |
| 🚨 extract_sec_user_id_batch | bulk_extraction | bulk identifier extractor |
| 🚨 extract_unique_id_batch | bulk_extraction | bulk identifier extractor |
| 🚨 encrypt_login | login_crypto | login request crypto helper; account-credential adjacent |

---

> 本文件是原子化层(Atomic Layer)的索引。每个端点封装为一个"原子",具备标准化输入/输出/幂等性/链路角色声明。
> Agent 可按 atom_id 快速定位端点,无需全文读取 reference 详情。

## 原子映射表

### 视频(video)

| atom_id | endpoint_id | file | d

The consent gate requires Chinese-language confirmation without offering a language choice.

Medium
Category
Output Handling
Confidence
93% confidence
Finding

Every accepted confirmation is in Chinese, while English “yes” is explicitly rejected and execution is blocked until an accepted confirmation is provided. This imposes a language requirement without user opt-in; the cited Chinese laws do not explain why consent cannot be expressed in the user's language.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The installation command can download and execute an unpinned version of the skillhub npm package, making installation depend on mutable registry content. Although this is not an MCP server reference as the detector claims, a compromised future release could execute code with the installing user's permissions.

Content

No source excerpt is available for this finding.

The compliance gate requires Chinese-language consent without offering a language choice.

Medium
Category
Output Handling
Confidence
96% confidence
Finding

The accepted confirmations are exclusively Chinese phrases, while English responses such as "yes" are explicitly rejected. This forces users to confirm in a specific language without opt-in or a documented justification, and the same restriction is reused for subsequent risk confirmations.

Content

No source excerpt is available for this finding.

The analytics skill instructs the agent to upgrade its installed implementation, including a forced upgrade, in response to API errors.

Medium
Category
Excessive Agency
Confidence
91% confidence
Finding

The troubleshooting guidance recommends replacing the installed skill, including a forced upgrade, based on API error responses rather than an independently authorized maintenance decision. If followed automatically, an upstream failure or manipulated response could cause installation of mutable, unreviewed instructions into an agent with access to API credentials and tools.

Content

No source excerpt is available for this finding.

Direct video-CDN requests contradict the manifest's API-host-only guarantee.

Medium
Category
Excessive Agency
Confidence
94% confidence
Finding

The instructions direct the agent to contact response-provided CDN URLs and attach a token, despite the declared API-host-only boundary. Video retrieval is relevant to this skill, but no CDN destination or redirect validation is specified; a malicious or compromised response could cause unintended outbound requests and disclosure of the CDN token.

Content

No source excerpt is available for this finding.

Generic validation phrases can activate a video-metrics recipe outside its intended domain.

Medium
Category
Trigger Abuse
Confidence
95% confidence
Finding

L052 defines “交叉验证” (cross-validation) and “数据验证” (data validation) as triggers for a video fake-traffic verification recipe. These phrases are common in unrelated statistics and data-quality requests, and the matching instructions at L004–L007 do not require video-specific intent before invocation.

Content

No source excerpt is available for this finding.

Confirmation rules require Chinese-language consent without offering a language choice.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

All accepted confirmation phrases are Chinese, while English confirmations such as "yes" are explicitly rejected. This requires users to confirm in Chinese without an opt-in or a documented reason that consent must use that language.

Content

No source excerpt is available for this finding.

L052 mandates a fixed Chinese-language reply without allowing for the user's language preference.

Low
Category
Output Handling
Confidence
90% confidence
Finding

The instruction tells the agent to reply directly with a specified Chinese sentence whenever an unsupported capability is requested. Unlike merely documenting the skill in Chinese, this prescribes user-facing output without a language choice or a justified locale restriction.

Content

No source excerpt is available for this finding.