Back to skill

Security audit

01 Product Selection

Security checks for vulnerabilities and agentic risk

Overview

This is a Chinese community group-buying product-selection knowledge skill with adjacent operational advice, but it does not install code, access data, persist, or perform actions automatically.

Installers should treat this as business guidance, not an autonomous operations tool. Review the Chinese-only content and manually validate any recommendations involving delisting, refunds, supplier blacklists, legal escalation, pricing formulas, or logistics changes before applying them to real users or suppliers.

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

Note
Location
SKILL.md:421
Finding

Unvalidated pricing rates can trigger division by zero

Content
View full analysis
Remediation
View remediation
= 1 for rate in rates): raise ValueError("Each rate must be greater than or equal to 0 and less than 1.") target_denominator = 1 - total_cost_rate - target_gross_margin breakeven_denominator = 1 - total_cost_rate if target_denominator <= 0: raise ValueError("Combined cost rates and target margin must be less than 1.") if breakeven_denominator <= 0: raise ValueError("Combined cost rates must be less than 1.") min_selling_price = cost_price / target_denominator recommended_price_b = cost_price / breakeven_denominator ``` ]]>

T09 · Insecure Skill Coding Practices

Note
Location
SKILL.md:599
Finding

Empty or malformed competitor-price input can terminate the comparison tool

Content
View full analysis
Remediation
View remediation
0 for price in all_prices ): return {"error": "All prices must be finite numbers greater than zero."} avg_price = statistics.mean(all_prices) min_price = min(all_prices) max_price = max(all_prices) ``` Example interactive validation: ```python market_input = input( "Market price (supermarket price, press Enter to skip): " ).strip() if market_input: try: market_price = float(market_input) if not math.isfinite(market_price) or market_price <= 0: raise ValueError except ValueError: print("Market price must be a finite number greater than zero.") return else: market_price = None ``` ]]>
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 (5)

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest says the skill does not include 供应链执行 or 仓储物流, but the body contains concrete operational guidance such as supplier blacklisting workflows, immediate delisting/refund/legal escalation steps, and loss-reduction plans involving procurement cadence, packaging, routing, storage, and pickup handling. Those are not just implementation details of product selection; they extend into supplier management and logistics execution contrary to the declared scope boundary.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill description and all instructional content are written entirely in Chinese, with no indication that users may choose another language or locale. Under the stated policy, a skill that effectively requires one language without opt-in or documented justification is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The boundary note states the module focuses on what to sell and pricing, and that supply chain plus warehousing/logistics are separate and non-overlapping. Later content contradicts this by detailing cold storage loss sources, packaging changes, center-warehouse-to-group-leader routing changes, daily procurement cadence, and sorting-time requirements, which are directly logistics/warehouse operations.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

Although introduced as 选品风控, this section gives actionable procedures for platform-wide delisting, notifying purchasers, refunds, public reporting, permanent blacklists, and legal escalation. Those activities are operational governance and supplier-management execution, not merely product-selection strategy or pricing support as described in the manifest.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This case does not just analyze selection outcomes; it prescribes operational interventions such as daily replenishment, packaging redesign, direct-shipping route changes, storage controls, and group-leader sorting deadlines. That is logistics and fulfillment optimization beyond the declared scope of a product-selection skill.

Content

No source excerpt is available for this finding.