Market Analysis Engine

Client handoff for a general market-analysis engine with a gold-first workflow.

Start here

• Client guide: docs/client-guide-grounded-gold.md
• Default workflow: grounded-gold
• Project guidance: .claude/CLAUDE.md

For ordinary gold-market questions, use grounded-gold. It fetches current gold-market facts where possible, labels freshness honestly, separates facts from interpretation, and ends gold reads with an evidence ledger.

Claude Code auto-routing is enabled through .claude/settings.json: a UserPromptSubmit hook adds grounded-gold context for gold-market and default client-market prompts. Explicit local-engine requests still use the broader engine.

Safety boundary

This system is decision support only.

• It is not a trading bot.
• It has no live trading workflow.
• It does not place orders.
• It is not connected to a broker for execution.
• It does not provide investment advice.
• It does not promise profit.

What to ask

• "What is gold doing today?"
• "Why is gold up or down?"
• "What are the bullish and bearish cases?"
• "What data is stale or missing?"
• "Show the evidence behind the read."

Use the broader engine only for explicit local research tasks: historical data ingest, intraday candidate setups, data-quality review, backtests, risk controls, and verification.

Verification

Run the project checks from the repo root:

.\.venv\Scripts\python.exe -m pytest
.\.venv\Scripts\python.exe -m ruff check .

Expected handoff state: all tests pass and ruff reports no findings.