Back to skill

Security audit

什么值得买

Security checks for vulnerabilities and agentic risk

Overview

This skill is a shopping-advice helper that compares public prices and gives purchase recommendations, with no code execution, persistence, account access, or payment authority in the inspected artifacts.

Install this if you want decisive Chinese-language shopping advice based on public listings and promotion comparisons. Treat its recommendations as buying advice, not proof of exact live cart prices, and keep control of any account login, coupon use, checkout, or payment yourself.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The description says to use the skill whenever the user gives a product name or category and wants a direct answer on what is cheapest, most worth buying, or most discounted. This broad condition lacks explicit exclusions or negative examples, so it could overlap with many ordinary product-comparison requests and cause unintended invocation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The instruction "This skill should sound like a decisive Chinese shopping advisor" imposes a specific language and locale style. The file repeatedly reinforces Chinese-only phrasing and Chinese internet shopping tone without offering user choice or documenting a justified regional restriction.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 429)May include surrounding context.

md
- clearly mark assumptions when exact cart price is unavailable

Stop before:
- logging into user accounts without consent
- adding irreversible orders
- claiming coupons or submitting payment

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The default prompt hard-codes that the assistant should respond in Chinese regardless of the user's language preference. This can override user intent, reduce usability, and in some contexts cause misunderstandings about purchasing advice, though it is not a direct security compromise. In this shopping skill context, the issue is less dangerous than prompt injection or data exfiltration, but it is still an instruction-layer policy problem because it removes user choice without opt-in.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The file labels several phrases as "Typical triggers" but does not state whether these are exhaustive, nor does it provide contrasting examples of similar requests that should not activate the skill. That makes the boundary between this skill and general shopping assistance unclear.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The manifest description and keyword set are entirely in Chinese for the skill identity and usage context, with no indication that users may interact in another language or choose a locale. Under the policy for natural-language constraints, this can be a locale/language restriction that is not explicitly optional or justified.

Content

No source excerpt is available for this finding.