Back to skill

Security audit

Precise T Trading

Security checks for vulnerabilities and agentic risk

Overview

This trading skill is mostly purpose-aligned, but it needs Review because it gives actionable financial recommendations while using unsafe and under-disclosed network, filesystem, and privilege practices.

Review carefully before installing or using with real money. Run it only as a normal user, prefer a virtual environment with pinned dependencies, avoid the HTML dashboard unless the JSONP/innerHTML issues are fixed, and treat all trading outputs as educational signals requiring independent verification.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (5)

T03 · Remote Payload Retrieval and Execution

Error
Location
scripts/dashboard.html:158
Finding

Remote JavaScript Payload Execution Through JSONP

Content
View full analysis
{ const script = document.createElement('script'); const callbackName = 'stockCallback_' + Date.now(); window[callbackName] = function(data) { resolve(data); delete window[callbackName]; document.body.removeChild(script); }; script.src = `https://qt.gtimg.cn/q=${code}&callback=${callbackName}`; script.onerror = () => resolve(null); document.body.appendChild(script); ``` ### Technical Analysis The dashboard creates a `
Remediation
View remediation
` elements. 2. Retrieve quote data as inert JSON using `fetch()` from an endpoint that explicitly supports CORS. 3. If the upstream API does not provide JSON with CORS, use a small trusted local backend that fetches and parses the quote response, validates it, and returns a strict JSON schema. 4. Validate ticker codes against a restrictive pattern such as `^(sh|sz)[0-9]{6}$`. 5. Validate every returned field's type, range, and maximum length before rendering. 6. Add a restrictive Content Security Policy, for example limiting `script-src` to `'self'` and avoiding `unsafe-inline`. 7. If remote scripts are unavoidable, proxy and pin reviewed content rather than executing a mutable third-party response directly. ]]>

T09 · Insecure Skill Coding Practices

Error
Location
scripts/dashboard.html:205
Finding

DOM Injection Through Unescaped Remote Quote Fields

Content
View full analysis
${stock.name}
${stock.code}
数据获取失败
`; } const isUp = data.changePct >= 0; const priceClass = isUp ? 'price-up' : 'price-down'; const badgeClass = isUp ? 'change-up' : 'change-down'; const changeSymbol = isUp ? '+' : ''; return `
${data.name}
${stock.code}
¥${data.price.toFixed(2)}
${changeSymbol}${data.change.toFixed(2)} (${changeSymbol}${data.changePct.toFixed(2)}%)
`; } async function refreshData() { const grid = document.getElementById('stockGrid'); grid.innerHTML = '
加载中...
'; const cards = []; for (const stock of stocks) { const data = await fetchStockData(stock.code); cards.push(createStockCard(stock, data)); } grid.innerHTML = cards.join(''); } ``` ### Technical Analysis `data.name` originates from the remote quote re ...[truncated 1611 chars]
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
requirements.txt:1
Finding

Unpinned and Globally Installed Third-Party Dependencies

Content
View full analysis
=1.24.0 scipy>=1.10.0 requests>=2.28.0 colorama>=0.4.6 ``` `SKILL.md`: ```text install: pip install numpy scipy requests colorama ``` ```text | "Module not found" | Run `pip install -r requirements.txt` | ``` `PUBLISH_GUIDE.md` and `QUICK_START.md`: ```bash npm install -g clawhub ``` ### Technical Analysis The Python requirements specify only minimum versions and do not include hashes or upper bounds. The direct installation instruction in `SKILL.md` does not specify versions at all. Future package releases can therefore be selected without having been reviewed with this Skill. The publishing documentation also recommends globally installing an unpinned npm package. Global installation expands the affected scope and may expose the user's command environment to package lifecycle scripts and future package behavior. No evidence indicates that the named dependencies are malicious or that dependency confusion is currently being exploited. The weakness is that package installation trusts mutable registry state and unreviewed future releases. ### Attack Path 1. A user follows the documented pip or npm installation command. 2. The package manager resolves the latest release satisfying the open-ended constraint. 3. A compromised package account, registry incident, malicious update, or dependency-chain compromise introduces hostile installation or runtime code. 4. The package manager downloads and executes the affected package code with the user's privileges. 5. For a global npm installation, the malicious package may also affect commands and files outside this project environment. ### Impact Assessment A compromised dependency c ...[truncated 512 chars]
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
SKILL.md:223
Finding

Unnecessary Recommendation to Run the Skill as Administrator

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Note
Location
scripts/stock_monitor.py:84
Finding

Undeclared Write to a Developer-Specific Absolute Filesystem Path

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (22)

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The skill claims advanced quantitative trading logic, risk management, and optimization that the associated behavior reportedly does not implement. This is dangerous because users or agent orchestrators may rely on false safety and decision-quality claims in a high-risk financial context, leading to harmful actions based on misleading capability descriptions.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The guide tells users to execute local publish scripts that may install global tooling, perform network access, and publish content, but it does not clearly disclose those side effects before execution. This creates a social-engineering and supply-chain risk because users are encouraged to run automation without understanding what system changes or outbound actions will occur.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
84% confidence
Finding

The skill advertises network access and local writes but does not declare an explicit tool scope such as permissions or allowed-tools. In agent environments, that ambiguity can cause the host to grant broader capabilities than users expect, reducing transparency and increasing the chance of unintended network requests or filesystem modification.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The invocation guidance uses broad triggers like general requests about trading strategies, stock analysis, risk management, and monitoring without clear boundaries. In an agentic setting, that can cause over-triggering on ordinary financial conversation, leading the skill to activate unexpectedly, make network calls, write logs, or present authoritative trading advice when the user did not explicitly request this tool.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The example triggers are short, natural conversational phrases that could easily appear in normal dialogue, making accidental invocation more likely. In this financial context, unintended activation is more dangerous because it may generate trading recommendations, open dashboards, or start monitoring workflows with implied authority and side effects.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The manifest describes a professional T+0 intraday trading system with quantitative decision optimization, position sizing, and risk control. In this file, the code only fetches Tencent quote data and renders a simple monitoring dashboard with periodic refresh; there is no trading logic, Bayesian inference, Kelly criterion, VaR calculation, or automated decision/risk-control behavior present.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The top-level documentation says the script automatically checks stocks every hour, which implies continuous monitoring. However, the actual entrypoint calls monitor_once() once and leaves the continuous main() loop commented out, so the default behavior contradicts the stated intent.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The file’s natural-language description and user-facing output are entirely in Chinese, including the docstring and console messages. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is explicitly justified, which is not present here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The module title, descriptions, and all user-facing terminal output are written in Chinese, and there is no mechanism for user opt-in or language selection. The stated policy requires flagging language or locale constraints unless the skill offers a choice or clearly documents a justified region-specific limitation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The script outputs explicit GO/WAIT trading recommendations, price bands, position sizing, and expected profit without any warning that the content is informational, model-based, and subject to market/data error. In a financial-decision context, omission of such a warning increases the chance that users will over-trust the tool and act on imperfect or simulated assumptions.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The displayed buy/sell ranges are derived from a hardcoded current price (4.06) instead of the fetched live quote, which can materially misstate actionable trading guidance while presenting the output as real-time analysis. In a trading skill, inconsistent decision inputs and displayed execution levels can directly mislead users into placing trades at inappropriate prices, increasing loss risk.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

This markdown file contains user-facing instructions and promotional content exclusively in Chinese, including headings, steps, and example copy. Under the policy rule, forcing a specific language without user opt-in can be a natural-language policy violation unless the locale restriction is explicitly documented and justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

A language or locale policy concern exists when instructions force a specific language without user opt-in. This file presents all operational guidance exclusively in Chinese and does not indicate multilingual availability or that the locale is intentionally limited.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
96% confidence
Finding

The dependency is specified with a lower bound only, so builds may resolve to different versions over time. This weakens supply-chain control and can allow accidental installation of vulnerable or incompatible releases, though by itself it does not prove a compromise.

Content

Scanner excerpt · requirements.txt (reported line 1)May include surrounding context.

text
numpy>=1.24.0
scipy>=1.10.0
requests>=2.28.0
colorama>=0.4.6

Unverifiable Dependency: numpy has 16 known advisory(ies) (CVE-2014-1859 (Numpy arbitrary file write via symlink attack); CVE-2021-41495 (NumPy NULL Pointer Dereference); CVE-2021-33430 (NumPy Buffer Overflow (Disputed)) +13 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
90% confidence
Finding

numpy has known advisories, and because the manifest does not pin an exact version, there is no way to verify whether installations will avoid affected releases. This creates uncertainty in dependency risk posture rather than proving active exploitation.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
96% confidence
Finding

The scipy requirement is unpinned, which makes installations non-reproducible and leaves the environment exposed to whatever newest matching version is available at install time. In a trading skill, inconsistent numerical library behavior can also increase operational risk alongside supply-chain exposure.

Content

Scanner excerpt · requirements.txt (reported line 2)May include surrounding context.

text
numpy>=1.24.0
scipy>=1.10.0
requests>=2.28.0
colorama>=0.4.6

Unverifiable Dependency: scipy has 4 known advisory(ies) (CVE-2013-4251 (SciPy creates insecure temporary directories); CVE-2013-4251 (The scipy.weave component in SciPy before 0.12.1 creates insecure temporary dire); CVE-2023-25399 (A refcounting issue which leads to potential memory leak was discovered in scipy) +1 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
89% confidence
Finding

scipy has historical advisories, and the lack of version pinning prevents verification that deployed environments use a safe release. In a quantitative trading context, this is primarily a software supply-chain hygiene issue rather than a direct exploit path from the requirements file alone.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
98% confidence
Finding

An unpinned requests dependency is more concerning because it is network-facing and commonly handles redirects, authentication, and TLS-related behavior. Allowing any version above a minimum increases the chance of pulling a release with newly introduced issues or unresolved known vulnerabilities.

Content

Scanner excerpt · requirements.txt (reported line 3)May include surrounding context.

text
numpy>=1.24.0
scipy>=1.10.0
requests>=2.28.0
colorama>=0.4.6

Unverifiable Dependency: requests has 16 known advisory(ies) (CVE-2014-1830 (Exposure of Sensitive Information to an Unauthorized Actor in Requests); CVE-2024-47081 (Requests vulnerable to .netrc credentials leak via malicious URLs); CVE-2024-35195 (Requests `Session` object does not verify requests after making first request wi) +13 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
95% confidence
Finding

requests has multiple security advisories, including issues relevant to credential handling and network interactions, and the unpinned requirement makes it impossible to confirm a safe installed version. Because this skill explicitly uses real-time quote retrieval, a network library with unverifiable versioning is more consequential here than a purely local utility.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
93% confidence
Finding

colorama is also unpinned, so the environment is not fully reproducible and may drift over time. The direct security impact is comparatively small because it is typically a terminal-color utility, but it still represents avoidable supply-chain uncertainty.

Content

Scanner excerpt · requirements.txt (reported line 4)May include surrounding context.

text
numpy>=1.24.0
scipy>=1.10.0
requests>=2.28.0
colorama>=0.4.6

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

The document declares lang="zh-CN", and the UI text and timestamp formatting are consistently fixed to Simplified Chinese. Under the policy, forcing a specific language or locale without user opt-in is a natural-language policy concern unless the restriction is clearly justified or configurable.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

toLocaleString('zh-CN') forces date/time presentation in a specific locale. This creates a locale policy issue because the user is not given a choice and the file does not explain why Chinese formatting is mandatory.

Content

No source excerpt is available for this finding.