AI-powered market intelligence framework — fetch, analyze, and deliver daily market briefings. Exposes market intelligence tools to AI assistants via MCP protocol.
MarketBrief defines 7 tools with complete input schemas and descriptions. Naming follows verb_noun convention (generate_report, fetch_market_data, analyze_regime). Descriptions are substantive (100-250 chars), explaining WHAT each tool does and WHEN to use it. However, output schemas are entirely undocumented, LLMs cannot infer what fields to expect from responses. Parameter descriptions are present but lack format constraints (e.g., 'date' accepts ISO format but no pattern validation stated). No error handling guidance, no recovery paths, and no mention of API key requirements in parameter descriptions. Security concern: generate_report requires ANTHROPIC_API_KEY but does not document secret injection pattern. analyze_breadth has empty inputSchema properties, suggesting incomplete definition.
Run the market breadth analyzer — measures market participation using advance/decline ratios, new highs/lows, moving average crossovers, and divergence signals.
Run the macro regime detector — identifies structural market regime shifts using cross-asset ratios (yield curve, credit conditions, equity-bond correlation, sector rotation, concentration). Returns regime classification and confidence scores.
Fetch economic calendar events from Forex Factory, MyFXBook, and FRED. Returns scheduled data releases, central bank decisions, and other market-moving events with times, impact levels, and actual values.
Fetch ETF flow and AUM data for BTC, ETH, and Gold spot ETFs. Uses SoSoValue API (primary) with RSS fallback.
Fetch current market snapshot — equities, commodities, FX, rates, volatility, and crypto prices. Uses Yahoo Finance (primary) and Stooq (fallback). No API key required.
Fetch and aggregate news from 40+ RSS feeds covering macro, markets, crypto, AI/tech, geopolitics, and government sources. Returns deduplicated items with source, title, URL, and timestamp.
Output schemas completely undocumented. LLMs cannot infer response structure, field names, or types. Breaks downstream tool chaining and forces agents to guess at response format.
analyze_breadth has empty inputSchema.properties ({}), indicating incomplete tool definition. No parameters documented despite tool accepting configuration.
generate_report requires ANTHROPIC_API_KEY but does not document secret injection pattern or clarify whether key is passed as parameter (security risk) or injected server-side (correct).
Parameter format constraints missing. 'date' parameter in fetch_calendar states 'ISO format, e.g. 2026-04-07' but lacks JSON Schema pattern or minLength/maxLength. LLMs may pass invalid dates.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 55 | <=2025-11-25 | v2 |
Generate a full AI-powered market briefing report. Fetches market data, news, and calendar, then uses Claude to produce structured analysis with tagline, today's focus, 4-issue analysis, positioning table, news digest, and economic calendar. Requires ANTHROPIC_API_KEY.
No error handling guidance. Tools do not document what happens on API failures (e.g., Yahoo Finance unavailable, API rate limits, missing data). LLMs have no recovery path.