mcpField
M
mql5-docs-research

mql5-docs-research

github.com/tmaru-engDevelopment0144 installsUpdated 40w ago
Repository

About mql5-docs-research

MQL5 Docs Research (JA/EN) is a documentation-based research skill for people troubleshooting or changing MQL5 code. It consults only the official Japanese and English documentation and structures findings with links and follow-up checks.

Key features

  1. 01Official Japanese and English docs only
  2. 02Collects error and code context
  3. 03Searches and verifies API details
  4. 04Guidance with doc links
  5. 05Follow-up checks for ambiguity

Use cases

  1. 01Investigating an MQL5 compiler error
  2. 02Checking an API signature or parameter
  3. 03Verifying documented return and error handling

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/tmaru-eng/strategy-bricks --skill mql5-docs-research

After installing, type/mql5-docs-researchin Claude Code to confirm it loaded.

Details

From the project README

MQL5 Docs Research (JA/EN)

Provide doc-based guidance when users hit errors while creating or modifying MQL5 code. Use only the official docs in Japanese and English.

Scope and Sources

Inputs to Collect

Ask for any missing context before research:

  • Exact error text and error code (if any)
  • File path and line/column numbers
  • MQL5 file type (.mq5 or .mqh)
  • Related API/class names (e.g., CTrade, MqlRates)
  • MetaTrader 5 build/version (if available)

Workflow

  1. Normalize the error or question into keywords:
    • Extract error code and symbols (function/class names)
    • Keep 2-5 keywords for searching
  2. Search Japanese and English docs:
    • Use site search in the docs pages
    • If no direct hit, search broader sections by keyword
  3. Open the most relevant doc pages and confirm:
    • Correct function signatures
    • Parameter types and constraints
    • Return values, error handling, and examples
  4. Produce a response in Japanese with:
    • Summary of the cause
    • Doc links (JA first, EN second)
    • Specific guidance aligned with doc wording
    • Next checks (if ambiguity remains)

Output Format (Markdown)

  • Summary (1-3 sentences)
  • Relevant Docs
    • JA:
    • EN:
  • Key Notes (bullets)
  • Next Checks (bullets, optional)

Guardrails

  • If the error cannot be mapped to docs, state that clearly and propose a best-effort hypothesis.
  • Never invent API signatures or behavior not documented in the official docs.
  • Keep the response concise and actionable.

Comments

Markdown supported