Back to skill

Security audit

07 After Sales Service

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed Chinese after-sales/customer-service reference skill with advisory calculators; it has scope and validation cautions but no hidden execution, persistence, or data access.

Use this skill as Chinese-language guidance and decision support, not as an automated authority. Validate inputs and require human/business approval before applying refunds, commission deductions, supplier penalties, blacklisting, delisting, or payment freezes.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:429
Finding

Unvalidated Numeric Inputs Permit Invalid Refund Calculations

Content
View full analysis
= 0 quantity_ordered >= 0 0 <= quantity_acceptable <= quantity_received <= quantity_ordered ``` If `quantity_received` or `quantity_acceptable` exceeds `quantity_ordered`, the subtraction produces a negative quantity and therefore a negative refund. Negative prices or quantities can also produce invalid totals. Conversely, inconsistent values may result in excessive or otherwise incorrect refund recommendations. The interactive wrapper catches only conversion errors. It does not reject successfully parsed but invalid values. The function itself also exposes no validation boundary, so direct callers can supply arbitrary numeric arguments. Using binary floating-point values for currency additionally creates a risk of rounding discrepancies, although the primary vulnerability is the absence of range and consistency validation. ### Attack Path 1. An attacker or malformed upstream record supplies inconsistent values, such as: ```python calculate_refund( product_price=25.0, quantity_ordered=2, quantity_received=5, quantity_acceptable=0, problem_type="shortage", is_platform_fault=False, ) ``` 2. The function calculates: ```text 25 × (2 - 5) = -75 ``` 3. The negative amount is returned as the recommend ...[truncated 944 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Note
Location
SKILL.md:586
Finding

Zero Aggregate Order Count Causes KPI Analysis Denial of Service

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest description is written only in Chinese, and the rest of the markdown continues in Chinese-only instructions and examples. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is explicitly documented and justified, which is not present here.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The boundary note explicitly states the module focuses on refunds/returns, damage compensation, and complaint handling, and that supply-chain procurement is covered by a separate module with independent boundaries. Later code contradicts that by implementing a '售后-供应链联动触发器' that recommends supplier blacklisting, fines, full-product takedowns, and payment freezes.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The manifest and boundary text frame this skill as focused on after-sales handling, complaint response, customer-service operations, and policy comparison, explicitly saying supply-chain procurement belongs to another independent module. Tool 4 goes further by generating supplier penalties, blacklist decisions, product delisting, payment freezing, and procurement-grade downgrades, which are operational supply-chain governance actions rather than customer-service tooling.

Content

No source excerpt is available for this finding.