Back to skill

Security audit

stock-monitor-a

Security checks for vulnerabilities and agentic risk

Overview

This stock-monitoring skill mostly matches its stated purpose, but it schedules automatic checks and routes alerts to a hardcoded external Xiaoyi delivery target that is not clearly user-controlled.

Review and change stock_config.json before installing, especially the stock list, thresholds, schedule, delivery_channel, and delivery_target. Install only if you are comfortable with recurring market checks, local retention of price and alert logs, and alert details being sent through the configured Xiaoyi channel.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (9)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding

The skill declares operational capabilities that imply file read/write and network access, but the manifest does not disclose any explicit tool scope or permissions boundary. That creates a security and governance gap: users and the hosting platform cannot clearly evaluate or constrain what the skill may access, which is especially relevant here because it fetches market data, writes persistent logs, and pushes alerts externally.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill documentation mentions immediate alert delivery through the Xiaoyi channel, but it does not present a clear warning that alerts are sent to an external delivery path automatically. This matters because users may trigger or schedule the skill without realizing data about monitored securities and alert conditions will be transmitted off-platform in real time.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill states that check/query operations automatically record price logs, but the description does not prominently warn users that these actions create persistent records. Silent persistence can violate user expectations, create unnecessary data retention, and increase exposure if logs later include sensitive operational context, query history, or account-specific watchlists.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
82% confidence
Finding

The manifest uses Chinese-language user-facing metadata for the skill name and description, but does not state that the skill is region-specific or that users can choose their preferred language. This can violate language/locale policy when a skill implicitly forces a specific language without opt-in or documented justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

This manifest contains user-facing natural-language fields such as the skill name, description, cron job name, and notes exclusively in Chinese. Under the policy, forcing a specific language without offering a user language/locale choice or documenting a justified regional constraint is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The module docstring and all user-facing descriptions are written only in Chinese, and the interface later accepts Chinese command aliases while providing no language/locale opt-in. This creates a natural-language locale constraint that appears mandatory rather than optional or explicitly justified as region-specific policy.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest says the skill supports '盘前盘后自动跳过' and filtering invalid market data, implying those periods are silently skipped. In code, get_real_price returns None for zero/invalid quotes, but check_stocks turns that into a visible '获取行情失败' alert-like message instead of skipping execution for pre-/post-market conditions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The run() interface exposes command aliases and responses in Chinese alongside English tokens, but the returned help and normal responses remain Chinese-only with no mechanism for the user to choose locale. For users not expecting Chinese output, this effectively forces a language choice without consent.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

The provided manifest context states version 3.0 with '无异常静默' and '即时推送', while the in-code skill metadata is version 2.2.0 and describes only querying/checking/log retrieval. The run path also returns '✅ 所有指标正常' on no alerts rather than being silent, so the implemented behavior does not match the claimed manifest semantics.

Content

No source excerpt is available for this finding.