Back to skill

Security audit

散户大佬帮你诊断行情(反人性报告)

Security checks for vulnerabilities and agentic risk

Overview

The scanners report actionable trading guidance that conflicts with the skill’s stated diagnosis-only purpose, so the verdict is suspicious; some reference checks were incomplete.

Consider the risk of receiving actionable financial guidance presented as structural analysis before installing. The scan found no reported persistence or privilege issue, but it was marked partial because several local references could not be resolved, so the evidence is incomplete.

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
  • 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 (2)

The report workflow prescribes position reductions despite the manifest's explicit promise to provide structural diagnosis without trading instructions.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

The workflow requires position-reduction and stop-loss sections, and its default funding alert explicitly directs reducing exposure regardless of price. This contradicts the skill's diagnosis-only boundary and can cause the agent to deliver actionable trading guidance under an informational label; the disclaimer does not resolve that contradiction. The financial context increases the potential for monetary harm, although no automated trading execution is shown.

Content

No source excerpt is available for this finding.

The dashboard provides actionable trading and position-sizing instructions despite the manifest limiting it to structural diagnosis without buy/sell instructions.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The template couples explicit price and fund-flow thresholds with instructions to reduce exposure, and its summary also directs increasing positions or exiting in batches. These are actionable trading recommendations despite the dashboard's diagnosis-only statements, making it likely that generated reports will reproduce guidance outside the advertised boundary. The risk is misleading financial guidance and possible losses, not demonstrated unauthorized trade execution.

Content

No source excerpt is available for this finding.