# Bloom Data > Foundation skill for all Bloom investing skills. Load this first before running any other Bloom skill. Handles CLI setup, authentication checks, output conventions, and documents every available bloom command with its flags. Triggers: "set up bloom", "bloom CLI help", "check bloom install", "bloom commands" - Skill: `exiao/bloom-data` (Agent Skill) - Install (CLI): `npx skillmds@latest add exiao/bloom-data` - Raw SKILL.md: https://api.skillmd.com/api/skills/exiao/bloom-data/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: exiao (https://skillmd.com/u/exiao) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/exiao/bloom-data --- # bloom-data — Foundation > Every number comes from Bloom's financial data. Not a language model's training data. This skill is the shared foundation. Every other Bloom skill reads this first. It defines the rules, checks prerequisites, and lists every available CLI command. --- ## Step 0: Prerequisites Check Before running any bloom command, verify: ### 1. CLI is installed ```bash bloom --help # or, if not globally installed: npx @bloomai/cli --help ``` If that fails: ```bash npm install -g @bloomai/cli ``` ### 2. Authentication is set up Option A — API key (preferred for automation): ```bash export BLOOM_API_KEY=your_key_here ``` Option B — Interactive login: ```bash bloom auth login ``` ### 3. Test it works ```bash bloom info AAPL ``` You should see Apple's stock data. If you see an auth error, run `bloom auth login`. --- ## Output Rules These rules apply to every skill that builds on bloom-data: 1. **Always save to file.** Use `-o /tmp/bloom/.json` for every command. This keeps the raw JSON out of your context window. ```bash mkdir -p /tmp/bloom bloom info AAPL -o /tmp/bloom/research-info.json ``` 2. **Use `jq` to extract only what you need.** Don't load entire JSON files into context. Pull the specific fields: ```bash cat /tmp/bloom/research-info.json | jq '.price, .pe_ratio, .market_cap' ``` 3. **Never estimate or hallucinate financial data.** If the CLI returns no data, say so. Don't fill in numbers from training data. 4. **Cite source commands.** When presenting a key number, note where it came from: - "P/E 33.2 (via `bloom info AAPL`)" - "Revenue grew 12% YoY (via `bloom financials AAPL --type revenue`)" 5. **Create the output directory first:** ```bash mkdir -p /tmp/bloom ``` --- ## Available Commands Full reference for every `bloom` CLI command. ### `bloom info [TICKER2 ...]` Stock overview: price, market cap, P/E, analyst ratings, sector, summary. ```bash bloom info AAPL bloom info AAPL MSFT GOOGL # batch up to 10 tickers bloom info NVDA -o /tmp/bloom/nvda-info.json ``` Key output fields: `symbol`, `price`, `change_pct`, `market_cap`, `pe_ratio`, `sector`, `analyst_consensus`, `price_target`, `bottom_line` --- ### `bloom financials --type ` Historical financial metrics. Always specify `--type`. ```bash bloom financials AAPL --type revenue bloom financials AAPL --type operating_margin bloom financials AAPL --type free_cash_flow bloom financials AAPL --type net_earnings bloom financials AAPL --type revenue -o /tmp/bloom/aapl-revenue.json ``` Available types: `revenue`, `revenue_growth`, `operating_margin`, `net_earnings`, `net_earnings_growth`, `free_cash_flow`, `free_cash_flow_growth`, `stock_based_compensation`, `stock_based_compensation_growth` --- ### `bloom earnings ` Recent earnings history + next earnings date. ```bash bloom earnings AAPL bloom earnings AAPL -o /tmp/bloom/aapl-earnings.json ``` Key output fields: `next_earnings_date`, `recent_results`, `beat_miss_history` --- ### `bloom catalysts ` Upcoming events and catalysts that could move the stock. ```bash bloom catalysts AAPL bloom catalysts NVDA -o /tmp/bloom/nvda-catalysts.json ``` Key output fields: `catalysts` (list with `title`, `description`, `impact`) --- ### `bloom technicals ` Technical analysis: trend, momentum, support/resistance levels. ```bash bloom technicals AAPL bloom technicals TSLA -o /tmp/bloom/tsla-technicals.json ``` Key output fields: `trend`, `momentum`, `support`, `resistance`, `moving_averages`, `rsi` --- ### `bloom news [--limit N]` Recent news headlines for a ticker. ```bash bloom news AAPL bloom news AAPL --limit 5 bloom news AAPL MSFT --limit 3 -o /tmp/bloom/news.json ``` Default limit: 10. Recommended: `--limit 5` for research, `--limit 3` for briefings. Key output fields: list of articles with `headline`, `source`, `published_at`, `url`, `summary` --- ### `bloom sentiment` Market-wide sentiment: Fear & Greed index, AAII bull/bear survey, VIX. ```bash bloom sentiment bloom sentiment -o /tmp/bloom/sentiment.json ``` Key output fields: `fear_greed_index`, `fear_greed_label`, `aaii_bull`, `aaii_bear`, `vix` --- ### `bloom market --type [--limit N]` Market overview by category. ```bash bloom market --type major_indexes bloom market --type top_movers --limit 10 bloom market --type sector_performance bloom market --type major_indexes -o /tmp/bloom/indexes.json ``` Available types: `major_indexes`, `top_movers`, `sector_performance` Key output fields: - `major_indexes`: list with `name`, `price`, `change_pct`, `ytd_change` - `top_movers`: list with `ticker`, `change_pct`, `price`, `reason` - `sector_performance`: list with `sector`, `change_pct` --- ### `bloom portfolio ''` Analyze a portfolio. Input is a JSON object with ticker:weight pairs. ```bash bloom portfolio '{"AAPL": 0.4, "MSFT": 0.3, "GOOGL": 0.3}' bloom portfolio '{"NVDA": 0.5, "TSLA": 0.25, "AMD": 0.25}' -o /tmp/bloom/portfolio.json ``` Weights should sum to 1.0. Key output fields: `grade`, `concentration_risk`, `sector_breakdown`, `correlation`, `suggestions` --- ### `bloom screen "" ["" ...]` Screen for stocks matching criteria. ```bash bloom screen "market_cap > 10B" "pe_ratio < 20" "revenue_growth > 10" bloom screen "dividend_yield > 3" "payout_ratio < 60" bloom screen "market_cap > 10B" "pe_ratio < 20" -o /tmp/bloom/screener.json ``` Common filters: `market_cap > XB`, `pe_ratio < X`, `revenue_growth > X`, `dividend_yield > X`, `payout_ratio < X` Key output fields: list of matches with `ticker`, `pe_ratio`, `revenue_growth`, `market_cap` --- ### `bloom collections [--section
]` Curated stock lists by theme. ```bash bloom collections bloom collections --section VALUE bloom collections --section GROWTH bloom collections --section DIVIDEND bloom collections -o /tmp/bloom/collections.json ``` Available sections: `VALUE`, `GROWTH`, `DIVIDEND`, `MOMENTUM`, `DEFENSIVE` Key output fields: list of collection entries with `ticker`, `name`, `reason`, `score` when available --- ### `bloom ai-trades [--limit N]` Trades made by AI models in the AI Arena (real money, live portfolios). ```bash bloom ai-trades bloom ai-trades --limit 10 -o /tmp/bloom/ai-trades.json ``` Key output fields: `trades` (list with `model`, `ticker`, `direction`, `return_pct`, `date`) --- ### `bloom ai-portfolio` Current holdings of each AI model in the Arena. ```bash bloom ai-portfolio bloom ai-portfolio -o /tmp/bloom/ai-portfolio.json ``` Key output fields: `portfolios` (keyed by model name, with `holdings`, `total_return`, `cash_pct`) --- ### `bloom trades ""` Evaluate a trade thesis. Pass a natural language description. ```bash bloom trades "buy NVDA before next earnings because AI spending is accelerating" bloom trades "short TSLA because valuation is stretched" -o /tmp/bloom/check-eval.json ``` Key output fields: `verdict`, `bull_case`, `bear_case`, `risk_reward`, `supporting_data` --- ### `bloom position-size --bull --bear --conviction [--value ]` Calculate appropriate position size based on risk/reward. ```bash bloom position-size --bull 20 --bear 10 --conviction 7 bloom position-size --bull 30 --bear 15 --conviction 8 --value 100000 -o /tmp/bloom/sizing.json ``` - `--bull`: estimated upside % if thesis is right - `--bear`: estimated downside % if thesis is wrong - `--conviction`: score 1-10 - `--value`: optional portfolio value in dollars Key output fields: `position_pct`, `max_loss_pct`, `risk_reward_ratio`, `dollar_amount` (if value provided) --- ## Error Handling | Error | Fix | |-------|-----| | `command not found: bloom` | Run `npm install -g @bloomai/cli` | | `Authentication required` | Run `bloom auth login` or set `BLOOM_API_KEY` | | `Rate limit exceeded` | Wait 60 seconds, then retry | | `No data found for ` | Check ticker spelling; some OTC stocks aren't covered | | Empty JSON output | Check auth, retry once, then report no data available | --- ## Related Skills - `research` — Full stock analysis using bloom commands - `review` — Portfolio grading and suggestions - `discover` — Finding new opportunities - `check` — Evaluating a specific trade - `briefing` — Daily market update