Back to skill

Security audit

tradedaily

Security checks for vulnerabilities and agentic risk

Overview

This trading workflow skill is mostly purpose-aligned, but it handles sensitive portfolio data and uses unauthenticated HTTP market data in ways users should review before installing.

Review before installing. Treat the bundled portfolio data as sensitive, replace it with your own private local file or a synthetic example, and do not share generated reports without redaction. The risk monitor contacts an external quote service over HTTP and should not be relied on for trading decisions unless you replace it with an authenticated HTTPS data source and validate the results. The trading thresholds should be treated as examples, not financial advice.

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

Error
Location
scripts_files/risk_monitor.py:23
Finding

Unauthenticated HTTP Market Data Can Be Manipulated to Produce False Risk Alerts

Content
View full analysis
45: return { 'price': float(data[3]), 'change_pct': float(data[32]), 'name': data[1] } ``` ### Technical Analysis The risk monitor obtains security prices from a plaintext HTTP endpoint. HTTP provides neither transport encryption nor server authentication, so an attacker able to observe or modify network traffic can intercept the requested stock code and alter the returned price fields. The program directly trusts response fields `data[3]` and `data[32]`. These values control portfolio profit-and-loss calculations and the generation of stop-loss, take-profit, intraday-crash, and intraday-surge alerts. The implementation does not authenticate the response, verify the HTTP status, restrict redirects to HTTPS, or validate the returned quote against an independent source. Although the code checks the number of response fields, that structural check does not establish the integrity or authenticity of the data. ### Attack Path 1. A user invokes `risk_monitor.py` while connected through a network controlled or observable by an attacker, such as a hostile Wi-Fi access point, compromised router, or malicious proxy. 2. The script sends the portfolio stock code to `http://qt.gtimg.cn`. 3. The attacker intercepts the plaintext response. 4. The attacker substitutes forged values for the current price or daily percentage change. 5. The script parses the forged fields and treats them as authentic market data. 6. The risk calculations produce false results, potentially suppressing a genuine stop-loss warning or creating a fraudulent alert. 7. The user may make ...[truncated 735 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
config_files/portfolio.json:2
Finding

Real Personal Portfolio and Asset Data Is Distributed in Plaintext

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (13)

Tp4

High
Category
MCP Tool Poisoning
Confidence
93% confidence
Finding

The code partially matches the declared description only in one area: it generates opening, closing, and end-of-day report templates from a local portfolio file. However, the broader declared purpose describes a comprehensive trader task management system with real-time monitoring, risk alerts, task scheduling, trade log recording, and review functionality. None of those operational capabilities are present in this code chunk. The script is limited to formatting markdown text with mostly placeholder values and printing it to stdout. Therefore, the description materially overstates the implemented functionality, making this a mismatch.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This markdown file presents all user-facing instructions and report templates exclusively in Chinese, and does not indicate that language selection is optional or user-configurable. Under the policy rule for natural-language constraints, forcing a specific language without opt-in is a locale/language policy violation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This file gives highly prescriptive trading instructions such as fixed stop-loss, take-profit, liquidation, and position-sizing rules without any disclaimer that the material is not financial advice and may be unsuitable for a user's circumstances. In the context of a trading workflow skill, users may treat these rules as authoritative operational guidance, which can lead to financial losses or inappropriate risk-taking if applied blindly.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This code file contains natural-language instructions and descriptions that assume Chinese as the mandatory language for usage and operation. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation because users are not offered any locale or language choice.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The generated report content is hard-coded in Chinese across the report templates, and the script does not provide any user-controlled language or locale option. This creates a language policy issue because the skill imposes one language rather than offering an opt-in or clearly justified region-specific constraint.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

User-facing natural-language strings including the module description, help text, and runtime messages are exclusively in Chinese. Under the policy, forcing a specific language without opt-in or documented justification is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The report templates encourage handling portfolio positions, prices, and trading records without warning that these outputs may contain highly sensitive financial information. In a trading workflow, users may paste, save, or share these reports broadly, increasing the risk of unauthorized disclosure of holdings, strategy, and P&L.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The portfolio monitoring and configuration examples expose exact holdings, cash, and total asset values without any caution about sensitivity. In this skill context, that information can reveal trading strategy, concentration risk, and financial exposure, making accidental disclosure materially harmful.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
79% confidence
Finding

The manifest describes task scheduling, report generation, portfolio/risk monitoring, trading logs, and performance review. The documented workflow also includes 'Screen Stocks for Tomorrow' to prepare a watchlist, which is a distinct investment research/screening capability not declared in the manifest's features or use cases.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
87% confidence
Finding

The JSON content uses Chinese-only field values such as the portfolio name and stock/sector names, which imposes a specific language context without any documented user choice or locale opt-in. Under the language/locale policy, natural-language content should either offer a language choice or clearly justify the locale restriction.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

The entire skill document is written in Chinese and presents all task outputs, reporting steps, and workflow instructions in that language, with no indication that users may choose another language or that the skill is intentionally limited to a Chinese-speaking or China-market-only audience. This can violate language/locale policy when a skill forces one language without user opt-in.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
85% confidence
Finding

The script sends portfolio stock codes to an external third-party API over plain HTTP without any explicit disclosure or consent mechanism. While stock codes are not highly sensitive by themselves, outbound transmission can leak portfolio interests or trading activity patterns, and use of unencrypted transport allows interception or tampering of the response.

Content

No source excerpt is available for this finding.