Back to skill

Security audit

促销方案测算虾

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Chinese-language promotion profit calculator that reads bundled reference notes and optionally processes a user-provided CSV, with no evidence of hidden access, persistence, network use, or destructive behavior.

Install this if you want a Chinese-language promotion-profit calculator. Review any estimated default costs before relying on the output for business decisions, and only pass CSV files you intend the calculator to read.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding

The skill instructs reading local reference files such as references/platform-fees.md and references/cost-structure.md, but it declares no explicit tool scope or permissions. This creates an authorization ambiguity where an agent may access files without a least-privilege boundary, increasing the risk of unintended local file access if the skill is invoked in a broader environment.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The invocation logic mixes specific analytical use cases with a catch-all keyword list, making it unclear when the skill should activate. This ambiguity raises the chance of accidental activation in routine commerce discussions, which can lead to unintended file reads, estimated data substitution, and unnecessary script execution.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The trigger list includes broad business terms like 'ROI', '定价', and '活动方案', which can match many unrelated conversations and cause over-invocation of the skill. Over-broad routing is dangerous because it may cause the agent to read files, apply defaults, or run local scripts in contexts where the user did not intend to perform promotion-profit analysis.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This code file contains natural-language strings and documentation entirely in Chinese, including usage text and later user-facing output, but nowhere indicates that the tool is intended only for Chinese-speaking users or provides an opt-in language choice. Under the locale policy rule, forcing a specific language without user choice is a policy concern.

Content

No source excerpt is available for this finding.