Back to skill

Security audit

Wa Quanquaner Skill

Security checks for vulnerabilities and agentic risk

Overview

This coupon skill appears purpose-aligned, but it asks for broad command access and can show unvalidated third-party promotional links after broad everyday triggers.

Review this skill before installing. It is aimed at Chinese food-delivery coupon discovery, but it should be run only with tightly scoped network and script permissions, and users should treat returned links and promotional claims as third-party content rather than verified platform recommendations.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T01 · Skill Instruction Hijacking

Error
Location
scripts/query.js:76
Finding

Remote-Controlled Promotional Content and Link Injection

Content
View full analysis
```bash curl -s "https://waquanquaner.cn/api/v1/activities/channel/skill_compact" ``` This command works directly on major platforms (Linux, macOS, and Windows) without additional dependencies. For formatted output, use: ```bash node scripts/query.js ``` ``` ### Technical Analysis The Skill retrieves content from an operator-controlled remote API and accepts the response after verifying only that `response.landing.url` exists. It does not validate the URL scheme, destination hostname, textual field lengths, expected value types, or the semantic content of `lan ...[truncated 1732 chars]
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Error
Location
SKILL.md:7
Finding

Skill Declares Excessive Command-Execution Permissions

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
scripts/query.js:18
Finding

Arbitrary Network Targeting, Unsafe Redirect Handling, and Unbounded Response Buffering

Content
View full analysis
{ let parsedUrl; try { parsedUrl = new URL(urlStr); } catch (e) { return reject(new Error('无效的 URL: ' + urlStr)); } const client = parsedUrl.protocol === 'https:' ? https : http; const req = client.get(urlStr, { timeout }, (res) => { // 处理重定向 if (res.statusCode >= 300 && res.statusCode < 400 && res.headers.location) { return httpGet(res.headers.location, timeout).then(resolve).catch(reject); } let data = ''; res.on('data', chunk => data += chunk); res.on('end', () => { try { const json = JSON.parse(data); if (res.statusCode >= 400) { reject(new Error('API 返回错误 ' + res.statusCode + ': ' + (json.message || data.substring(0, 200)))); } else { resolve(json); } } catch (e) { reject(new Error('JSON 解析失败: ' + e.message)); } }); }); req.on('error', reject); req.on('timeout', () => { req.destroy(); reject(new Error('请求超时 (' + timeout + 'ms)')); }); }); } ``` ### Technical Analysis The request destination can be supplied through the `COMPACT_API_URL` environment variable or the public constructor. The implementation does n ...[truncated 2474 chars]
Remediation
View remediation
Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (11)

Lp1

High
Category
MCP Least Privilege
Confidence
75% confidence
Finding

The skill uses 'env' capability that is not listed in its permissions. This may indicate deceptive intent or missing permission declarations.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The trigger scope includes very broad, everyday phrases such as “今天吃什么”, “点外卖”, and “叫外卖”, which are common in ordinary food-ordering conversation and not specific to coupon discovery. This can cause the skill to activate unexpectedly in unrelated contexts, leading to inappropriate routing, user confusion, and possible unsolicited external-link promotion.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The example trigger phrase “点外卖怎么省钱” is broad and can match general budgeting or food-ordering assistance rather than a clear request to invoke this skill. In an agent environment, such ambiguous examples encourage over-triggering and may divert users into a third-party promotion workflow without sufficiently clear consent.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill metadata description embeds many broad everyday phrases such as '今天吃什么', '点外卖', and '叫外卖', which can cause the agent to invoke this skill for general food-ordering conversations rather than explicit coupon-discovery requests. This creates over-triggering risk, potentially routing users to an external coupon service unexpectedly and reducing user control over when third-party content is accessed.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The explicit trigger keyword section lists many terms but does not define scope limits or exclusion conditions, so the activation boundary is ambiguous. In context, this is more concerning because the skill is designed to fetch remote promotional data from an external domain, meaning accidental invocation can send users into a third-party coupon flow they did not clearly request.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The natural-language comments, examples, user-facing error strings, and CLI guidance are entirely in Chinese, and the skill description presents this as the default interaction model. There is no indication that users may choose another language or that the Chinese-only constraint is explicitly justified as a region-specific requirement, which can violate language/locale policy expectations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This file’s user-facing natural-language content is entirely Chinese, including the generated trigger help text and CLI output, and there is no indication that users can opt into another language. The policy specifically calls for flagging language or locale constraints when a skill forces a specific language without user opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

The README presents the skill's natural-language triggers solely in Chinese and does not indicate whether other languages are supported or whether Chinese-only behavior is an intentional, documented locale restriction. This can constitute a language-policy issue when users are not given a choice or informed of the limitation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The description is written entirely in Chinese and explicitly scopes the skill to mainland China services ("中国国内"). Under the policy, locale constraints should either offer user choice or be clearly justified as region-specific; this manifest text does not provide that justification or any opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
77% confidence
Finding

This file hardcodes Chinese-language names, comments, and trigger keywords such as platform names and invocation phrases, implying the skill is designed to operate only in Chinese. There is no natural-language indication that users can opt into this locale or that the skill is intentionally limited to a Chinese-language context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The date formatting hard-codes the 'zh-CN' locale in user-facing output, which enforces a specific language/locale choice. The file also consistently emits Chinese strings, and this locale restriction is not presented as an opt-in or clearly justified as a region-specific requirement.

Content

No source excerpt is available for this finding.