Skills quantall-mcp

📦

quantall-mcp

Name: quantall-mcp
Author: mifochen

Content revision r1 High Risk ⚙️ External commands🌐 Network access📁 Filesystem access

Analyze A-Share Strategies with QuantAll

Large A-share datasets make factor testing and strategy validation slow and inconsistent. QuantAll runs vectorized local analysis and returns measurable backtest, IC, screening, and visualization results.

Supports: Claude Codex Code(CC)

⚠️ 38 Poor

Stop for confirmation before installing.

Review the plan and obtain explicit user consent before changing files.

Install with my Agent

Copy this request to your Agent. It includes the canonical Skill page and manifest.

Agent request

Review the Skillstore skill "quantall-mcp" from https://skillstore.io/skills/mifochen-quantall-mcp.md and its manifest at https://skillstore.io/api/skills/mifochen-quantall-mcp/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.

Your Agent should still show its plan and request any confirmation required by the security policy.

Agent-readable resources

Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.

Markdown detail GET /skills/mifochen-quantall-mcp.md Signed manifest GET /api/skills/mifochen-quantall-mcp/manifest Signed lockfile GET /api/skills/mifochen-quantall-mcp/lockfile

Test it

Using "quantall-mcp". Test whether smaller market capitalization predicts stronger next-month returns.

Expected outcome:

Coverage summary with the tested universe, dates, and missing-data exclusions.
Factor report with IC direction, stability, quantile spread, and benchmark comparison.
Conclusion that separates measured evidence from interpretation and limitations.

Using "quantall-mcp". Backtest a moving-average crossover and compare close execution with next-open execution.

Expected outcome:

Side-by-side performance summary for both execution assumptions.
Return, drawdown, Sharpe ratio, win rate, turnover, and holding-period comparison.
Warnings about lookahead bias, data gaps, and sensitivity to parameters.

Using "quantall-mcp". Find low-valuation stocks with improving profitability and inspect their industry distribution.

Expected outcome:

Filtered stock set with each criterion stated clearly.
Industry distribution and concentration summary.
Data timestamp, field definitions, and excluded records.

Security Audit

High Risk

v1 • 7/24/2026 Open versioned report

All static pattern alerts are false positives caused by Markdown formatting, loopback addresses, scoped configuration writes, fixed subprocess arguments, or normal DuckDB access. Semantic review found high-risk sandbox-bypass guidance, unauditable Python execution, unverified remote installation, and SQL injection. Plaintext API token storage adds a medium-risk credential exposure.

Files scanned

3,745

Lines analyzed

Review items

False positives ignored

Confirmed security concerns (5)

High

SKILL.md:73

Sandbox Bypass Guidance

SKILL.md recommends `dangerouslyDisableSandbox: true` when installation conflicts occur. This removes a security boundary instead of resolving the underlying file operation.

The bypass setting is stated directly as an installation workaround on line 73.

High

SKILL.md:21-27 SKILL.md:197-207

AI-Generated Python Execution Is Not Auditable

The core workflow executes AI-authored Python snippets through QuantAll. Only prompt-level restrictions are visible, while the external engine enforcing them is absent.

The documentation explicitly describes local Python execution and behavioral restrictions. The repository does not contain the QuantAll engine implementation needed to verify isolation.

High

SKILL.md:44-59 SKILL.md:133-151

Remote Package Installation Without Integrity Verification

Installation fetches quantall without a version or hash. Upgrade guidance requests the latest release from a remote mirror, allowing unreviewed code changes.

The documented pip commands omit both a version constraint and hashes, and the upgrade command explicitly selects newer remote content.

High

scripts/UpdateStock_skill.py:1422-1429 scripts/UpdateStock_skill.py:1480-1487

SQL Injection in Stock Query Tools

MCP parameters for symbols and dates are interpolated directly into DuckDB SQL. Crafted values can change query structure and access or modify local data.

Both MCP tools concatenate unsanitized string parameters into executable SQL before passing it to DuckDB.

Medium

scripts/UpdateStock_skill.py:1503-1541 scripts/UpdateStock_skill.py:1553-1588

Plaintext Tushare API Token Storage

Update tools direct users to store a Tushare API token in API_tushare.txt. The script reads the secret without access controls or protected storage.

The tool documentation and implementation explicitly create and read API_tushare.txt as the credential source.

Risk Factors

⚙️ External commands (50)

scripts/UpdateStock_skill.py:1656 scripts/UpdateStock_skill.py:1665 scripts/UpdateStock_skill.py:1611 SKILL.md:27 SKILL.md:38 SKILL.md:45 SKILL.md:47 SKILL.md:50-69 SKILL.md:69-72 SKILL.md:72 SKILL.md:73 SKILL.md:74 SKILL.md:78-89 SKILL.md:89-95 SKILL.md:95-97 SKILL.md:97-103 SKILL.md:103 SKILL.md:105-109 SKILL.md:109 SKILL.md:111 SKILL.md:121-123 SKILL.md:123-127 SKILL.md:127-135 SKILL.md:135-137 SKILL.md:137-148 SKILL.md:148-150 SKILL.md:150 SKILL.md:151-152 SKILL.md:152 SKILL.md:162-164 SKILL.md:164-166 SKILL.md:166 SKILL.md:167 SKILL.md:175-179 SKILL.md:179-180 SKILL.md:180-181 SKILL.md:181-182 SKILL.md:182-183 SKILL.md:183-184 SKILL.md:184-185 SKILL.md:185-186 SKILL.md:186-187 SKILL.md:187-188 SKILL.md:188-189 SKILL.md:189-190 SKILL.md:190-191 SKILL.md:191-192 SKILL.md:192-193 SKILL.md:193-194 SKILL.md:194-195

🌐 Network access (11)

scripts/UpdateStock_skill.py:1625 scripts/UpdateStock_skill.py:1635 scripts/UpdateStock_skill.py:1683 scripts/UpdateStock_skill.py:1624 scripts/UpdateStock_skill.py:1625 scripts/UpdateStock_skill.py:1680 scripts/UpdateStock_skill.py:1683 SKILL.md:56 SKILL.md:81 SKILL.md:139 SKILL.md:81

📁 Filesystem access (24)

Audited by: codex

Share & cite this report

Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.

Open versioned report

Copy report link

https://skillstore.io/skills/mifochen-quantall-mcp/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown badge

[![Skillstore security assessment](https://skillstore.io/badges/skills/mifochen-quantall-mcp/security.svg)](https://skillstore.io/skills/mifochen-quantall-mcp?utm_source=security_passport_badge)

HTML badge

<a href="https://skillstore.io/skills/mifochen-quantall-mcp?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/mifochen-quantall-mcp/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Embed card

<iframe src="https://skillstore.io/embed/skills/mifochen-quantall-mcp.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

Academic citations (APA · BibTeX · CFF)

APA citation

mifochen. (2026). quantall-mcp security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/mifochen-quantall-mcp/audits/1

BibTeX citation

@techreport{mifochen-mifochen-quantall-mcp-2026,
  author = {mifochen},
  title = {quantall-mcp security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/mifochen-quantall-mcp/audits/1},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "quantall-mcp security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
  - name: "mifochen"
date-released: "2026-07-24"
url: "https://skillstore.io/skills/mifochen-quantall-mcp/audits/1"
identifiers:
  - type: other
    value: "skillstore:mifochen-quantall-mcp:audit:1"
    description: "Skillstore immutable audit report identifier"

Skillstore Score

Why this score Evidence Confidence: Medium

Architecture

Maintainability

Content

Community

Spec Compliance

What You Can Build

Validate a Trading Hypothesis

Backtest entry and exit rules across A-share history, then compare returns, drawdowns, and win rates.

Compare Candidate Factors

Measure IC, IR, and correlations for several factors before selecting complementary signals.

Maintain Local Market Data

Create or update a DuckDB dataset and configure the QuantAll engine to use it.

Try These Prompts

Inspect Available Data

Check the QuantAll service and list available fields. Summarize the date range, stock coverage, and missing data before analysis.

Backtest a Strategy

Backtest my stated entry and exit rules with close and next-open execution. Report returns, Sharpe ratio, drawdown, win rate, and assumptions.

Evaluate Multiple Factors

Compare these factors across the same period and universe: [factors]. Report IC, IR, stability, correlations, and redundant signals.

Run a Robust Research Workflow

Test [hypothesis] across [universe] and [period]. Check data availability, prevent lookahead bias, compare parameter ranges, and explain unstable results.

Best Practices

Obtain explicit approval before installing packages, changing MCP configuration, starting services, or creating files.
Call available_data before analysis and use only documented vectorized operations.
State the universe, dates, execution assumptions, data gaps, and lookahead controls in every result.

Avoid

Do not disable the sandbox or weaken host protections to complete installation.
Do not install unspecified package versions or upgrade remote code without review.
Do not run parallel tools, assume field names, or expose oversized detail outputs directly.