QVeris Financial Analyst Skills

Use this skill for financial document extraction, calculations, modeling, and JSON generation adapted from Financial Analyst Skills. Prefer QVeris filings and structured fundamentals over local PDF/model pipelines; local user-provided documents are fallback inputs only.

Source record:

Runtime Contract

• Use only qveris_finance.* CAP tools and QVERIS_API_KEY for default financial data.
• Resolve company, ticker, CIK, and filing identity with QVeris reference and filing tools.
• Accept dry_run, max_calls, max_age, and budget_note; if omitted in a natural-language request, default to dry_run=false, max_calls=12, max_age=P1D, and a conservative budget note, then echo those controls.
• Include qveris_trace for every extracted field, calculation input, model assumption, and JSON node.
• If a user supplies a local document, treat parsing as fallback, mark fallback_used: true, and never upload or expose private content unless explicitly requested by the user.
• Treat QVeris _meta.source_provider as provenance only; never call, request credentials for, or depend on those internal providers directly.
• Suppress analyst_target_price, target_price, price-objective, upside, buy/sell, and recommendation fields even if a QVeris payload contains them.
• Sanity-check entity, market, filing form, accession, fiscal period, and payload shape before using data; if a payload is stale, cross-period, metadata-only, truncated, or semantically mismatched, mark it in data_quality and missing_fields.

Workflows

1. Document classification/organization: filings_regulatory_metadata, filings_regulatory_raw, filings_structured_xbrl.
2. Financial data extraction: fundamentals_is, fundamentals_bs, fundamentals_cf, fundamentals_segment, filings_structured_xbrl.
3. Financial calculations: fundamentals_derived_ratios, mkt_l1_rt, mkt_bars_adjusted.
4. Financial modeling/DCF inputs: fundamentals_is, fundamentals_bs, fundamentals_cf, estimates_consensus, rates_govt_benchmark, fx_spot.
5. JSON model generator: emit trace-backed JSON; do not depend on local PDF extraction by default.

Output Requirements

• Use schemas/output.schema.json.
• Each extracted field must include source, period, unit, confidence, missing status, and trace.
• Calculations must distinguish QVeris-provided ratios from derived local calculations.
• If filings_structured_xbrl returns filing metadata rather than XBRL facts, do not label it structured XBRL evidence; fall back to QVeris fundamentals and mark XBRL line items missing.
• Align statement periods before modeling; if annual requests return quarterly or TTM-like cash flow data, keep that input separate and lower confidence.
• DCF output is assumptions and sensitivity only; no target price commitment.
• Include data_quality with status, stale fields, out-of-window events, and suppressed fields when applicable.
• End with: 不构成投资建议 / Not investment advice.

Prohibited Capabilities

Do not use local PDF/model pipelines as default financial data, EODHD, Yahoo, FMP, Alpha Vantage, Polygon, AkShare, Snowball, Sina, SEC scraping, Longbridge, FinViz, Alpaca, browser automation, cookies, login state, third-party API keys, automated trading, wallet/swap, buy/sell points, portfolio action instructions, or target price commitments.

References

• Read references/qveris-tool-map.md before choosing tool calls.
• Use fixtures/qveris/sample-output.json as the minimum output shape.