What does this skill do, and when should you use it?
This is a multi-source financial analysis skill for stock research, combining yfinance, SEC EDGAR, Finnhub, Alpha Vantage and 10+ other free APIs with 20+ RSS feeds. It supports three workflows — weekly portfolio review, daily opportunity scanner, and on-demand deep dive — producing Buy/Hold/Trim/Sell ratings, three entry levels, three price targets, and stop losses. Every API call flows through a resilient caller with rate-limit enforcement, logging, and automatic fallback chains. Scoring weights fundamentals at 40%, technicals 30%, and sentiment 30% into a composite 0-100 score.
Given a holdings list or ticker symbols, the skill: fetches price history, financial statements, analyst consensus, insider Form 4 filings, Congress trades, options chains, and news/Reddit/StockTwits sentiment; computes SMA/EMA, RSI, MACD, Bollinger Bands, Fibonacci retracement, ATR and more locally with pandas-ta; produces fundamental, technical, and sentiment scores weighted 40/30/30; and outputs ratings, 3 entry levels, 3 exit targets, stop loss, risk-reward ratio, catalysts, and risks. Calls go through api_caller.py, which checks rate limits and logs to ~/.financial-analysis/logs/, switching to the next API in the fallback chain on failure.
- An individual investor who wants a weekly (Sunday/Monday) Buy/Hold/Trim/Sell review of a PDF or CSV portfolio
- A swing trader running a daily pre-market scan of RSS and social buzz to surface top-5 portfolio adds and top-5 swing trades
- A user asking 'should I buy X?' who needs an analyst-grade deep dive on 1-10 tickers (~30 API calls each)
- A research-minded user tracking insider Form 4 filings and Congress trades, sourced from SEC EDGAR
- An investor gauging market mood via Reddit mention velocity, StockTwits sentiment, and news sentiment scores
How do you install this skill?
- Static review only; no code or tests were executed. Scores are based on source reading.
- Core data sources (Finnhub, Alpha Vantage, RapidAPI, etc.) are all overseas services; mainland-China network reachability is unverified and may seriously limit practical usability.
- The skill outputs buy/sell ratings and price targets — treat as data-driven reference, not advice; scoring weights are subjective and unbacktested. Always verify independently.
- Several key scripts (scoring.py, data_fetchers.py, etc.) were not included in the evidence; actual behavior may diverge from documentation.
- API keys are stored as plaintext JSON in the home directory — check file permissions; some free quotas are very low (e.g. Alpha Vantage 25/day).
- No version numbers or changelog; update and maintenance responsibility are unclear.
- Shell / CLI
- Network access
- Local filesystem
Pythonyfinancefeedparserpandaspandas-tarequests
Clone into the Claude skills directory: git clone <repo-url> ~/.claude/skills/financial-analyst. Then run ./setup.sh to create a venv and install dependencies (yfinance, feedparser, pandas, pandas-ta, requests), followed by python scripts/api_config.py init and status. Four sources (yfinance, SEC EDGAR, ApeWisdom, StockTwits) work with no API keys; Finnhub, Alpha Vantage, FMP, and Polygon require free keys — setup steps are in references/api_setup_guide.md (signup links are said to be in QUICKSTART.md, whose content is not included here).
git clone --depth 1 https://github.com/geogons/skill-financial-analyst.git ~/.claude/skills/skill-financial-analystGenerated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- review my portfolio
- deep dive on NVDA
Claude auto-loads the skill when you ask about stocks, portfolios, or market analysis. Three entry points: 1) Portfolio review: python scripts/run_portfolio_review.py AAPL:100:150.50 MSFT:50:380 (ticker:qty:cost format); 2) Daily scanner: python scripts/run_daily_scanner.py; 3) Deep dive: python scripts/run_deep_dive.py AAPL MSFT. Natural-language triggers also work, e.g. 'review my portfolio' or 'deep dive on NVDA'. After runs, check python scripts/usage_tracker.py daily for API usage reports.
What are this skill's strengths and limitations?
- Redundant data design: every category has a prioritized fallback chain (e.g. price history yfinance → Polygon → Alpha Vantage → FMP) with automatic failover
- Four core sources (yfinance, SEC EDGAR, ApeWisdom, StockTwits) need no API keys, lowering the barrier to first use
- Built-in usage tracker reports per-API free-limit consumption and recommends paid tiers (with prices) above 70% utilization
- Transparent, adjustable methodology: 40/30/30 weights and rating thresholds are documented and user-tunable
- Only recommends trades with R:R ≥ 2:1, with structured output (3 entries / 3 targets / stop / catalysts / risks)
- Relies on yfinance, an unofficial library the README itself flags for possible delays or outages
- Alpha Vantage free tier is only 25 calls/day; a deep dive uses ~30 calls per ticker, hitting limits fast
- Congress trades rely on Mboum (600/month free), with the fallback (Quiver) being paid
- No mention of a test suite; runtime scripts referenced in the README (setup.sh, run_*.py) are not detailed in SKILL.md
- Output is data-driven analysis, not financial advice — rating quality depends entirely on free-API data quality and freshness
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Financial Analyst Skill this page | 52 · Use with care | ★ 62 | 7mo ago | MIT |
| TradingView MCP AI Trading Intelligence | 50 · Use with care | ★ 5.1k | 11d ago | MIT |
| Deep Stock Analysis Workflow (UZI-Skill deep-analysis) | 50 · Use with care | ★ 7.2k | 1mo ago | MIT |
| Kronos Technical Analysis Skill | 44 · Not recommended | ★ 21 | 5mo ago | — |
| Daisy Financial Research | 55 · Use with care | ★ 29 | 5mo ago | — |
The source names no specific competitors. Versus calling these free APIs manually, its differentiation is automatic fallback chains, rate-limit management, and a unified scoring framework; versus paid research terminals (unnamed in the source), it trades data timeliness for zero subscription cost.
How did FollowSkills review this skill?
Files show least-privilege design: API keys stored in the user's home directory (~/.financial-analysis/api_keys.) and never overwritten by setup, with env-var alternative; no credential harvesting, covert exfiltration, or destructive defaults; external API queries carry only ticker-style lookups plus a disclaimer. Deducted for: automatic network calls to 14+ third-party APIs without per-call user confirmation, plaintext JSON key storage, unverified publisher, and rollback described only in docs.
A layered test suite (24 tests: offline/live/integration) is committed; error handling, degradation, rate limiting and fallback chains are documented and tested; offline tests run with zero configuration. Deducted for: static review without execution, no CI workflow proving tests pass, known instability of unofficial sources like yfinance, several referenced scripts (run_deep_dive.py etc.) not present in evidence, so key paths could not be statically reproduced.
Three use cases (weekly portfolio review, daily scanner, deep dive) have clear triggers, inputs and outputs; description carries rich trigger phrases and a not-financial-advice disclaimer. Deducted for: core function depends almost entirely on overseas services (Finnhub, Alpha Vantage, SEC EDGAR, RapidAPI) whose mainland-China reachability is doubtful and undisclosed; no Chinese-language support; capability boundaries (data delay, US-market-only coverage) thinly disclosed.
Well-layered docs (SKILL.md/QUICKSTART/README/setup guide), explicit MIT license, thorough install and dependency notes, common-error table and troubleshooting. Deducted for: no version number or changelog, unclear maintenance ownership ('Contributors'), unverified publisher, no declared update path.
Workflows are concrete (scoring formula, entry/exit structure, caching), output formats are well-defined including a data-source attribution table, and there is clear marginal value over manually aggregating many sources. Deducted for: static review cannot verify actual output correctness or usability, several APIs require key signup, tight quotas (Alpha Vantage 25/day) may prevent full deep-dive execution, and value claims are execution-unverified.
Auditable test code is committed (test_skill.py with 24 tests, test_enhanced_report.py); scoring methodology, fallback chains and rate limits are cross-checkable in docs; fact/inference separation is decent. Deducted for: no third-party execution evidence or CI results, tests never run in this review, multiple referenced scripts (data_fetchers.py, scoring.py) absent from evidence, and testimonial-style claims (e.g. '26 indicators') independently unverifiable.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →