finance-sentiment — скилл для Claude Code: сентимент акций

---
name: finance-sentiment
description: >
  Fetch normalized stock sentiment across Reddit, X.com, financial news, and
  Polymarket from the Adanos Finance API: buzz score, bullish percentage, mention or
  trade counts, and trend. Use this skill whenever the user asks how much a stock is
  being discussed or how hot it is on social platforms, how many Reddit mentions or
  Polymarket bets a company has, whether Reddit, X, and news agree, or wants to
  compare stock buzz and sentiment across tickers. Read-only: it never places trades
  or modifies anything.
---

# Finance Sentiment Skill

If your host has not already executed the shell probes below (for example, in Codex), run the commands inside the `!` wrappers with its shell tool before choosing a setup path. Use the actual output; if shell access is unavailable, report that limitation.

Fetches structured stock sentiment from the Adanos Finance API.

This skill is read-only. It is designed for research questions that are easier to answer with normalized sentiment signals than with raw social feeds.

Use it when the user wants:
- cross-source stock sentiment
- Reddit/X.com/news/Polymarket comparisons
- buzz, bullish percentage, mentions, trades, or trend
- a quick answer to "what is the market talking about?"

---

## Step 1: Ensure the API Key Is Available

**Current environment status:**

```bash
!`python3 - <<'PY'
import os
print("ADANOS_API_KEY_SET" if os.getenv("ADANOS_API_KEY") else "ADANOS_API_KEY_MISSING")
PY`
```

If `ADANOS_API_KEY_MISSING`, ask the user to set:

```bash
export ADANOS_API_KEY="sk_live_..."
```

Use the key via the `X-API-Key` header on all requests.

Base docs:

```text
https://api.adanos.org/docs
```

---

## Step 2: Identify What the User Needs

Match the request to the lightest endpoint that answers it.

| User Request | Endpoint Pattern | Notes |
|---|---|---|
| "How much are Reddit users talking about TSLA?" | `/reddit/stocks/v1/compare` | Use `mentions`, `buzz_score`, `bullish_pct`, `trend` |
| "How hot is NVDA on X.com?" | `/x/stocks/v1/compare` | Use `mentions`, `buzz_score`, `bullish_pct`, `trend` |
| "How many Polymarket bets are active on Microsoft?" | `/polymarket/stocks/v1/compare` | Use `trade_count`, `buzz_score`, `bullish_pct`, `trend` |
| "Compare sentiment on AMD vs NVDA" | compare endpoints for the requested sources | Batch tickers in one request |
| "Is Reddit aligned with X on META?" | Reddit compare + X compare | Compare `bullish_pct`, `buzz_score`, `trend` |
| "Give me a full sentiment snapshot for TSLA" | compare endpoints across Reddit, X.com, news, Polymarket | Synthesize cross-source view |
| "Go deeper on one ticker" | `/stock/{ticker}` detail endpoint | Use only when the user asks for expanded detail |

Default lookback:
- use `days=7` unless the user asks for another window

Ticker count:
- use compare endpoints for `1..10` tickers

---

## Step 3: Execute the Request

Use `curl` with `X-API-Key`. Prefer compare endpoints because they are compact and batch-friendly.

### Single-source examples

```bash
curl -s "https://api.adanos.org/reddit/stocks/v1/compare?tickers=TSLA&days=7" \
  -H "X-API-Key: $ADANOS_API_KEY"
```

```bash
curl -s "https://api.adanos.org/x/stocks/v1/compare?tickers=NVDA&days=7" \
  -H "X-API-Key: $ADANOS_API_KEY"
```

```bash
curl -s "https://api.adanos.org/polymarket/stocks/v1/compare?tickers=MSFT&days=7" \
  -H "X-API-Key: $ADANOS_API_KEY"
```

### Multi-source snapshot for one ticker

```bash
curl -s "https://api.adanos.org/reddit/stocks/v1/compare?tickers=TSLA&days=7" -H "X-API-Key: $ADANOS_API_KEY"
curl -s "https://api.adanos.org/x/stocks/v1/compare?tickers=TSLA&days=7" -H "X-API-Key: $ADANOS_API_KEY"
curl -s "https://api.adanos.org/news/stocks/v1/compare?tickers=TSLA&days=7" -H "X-API-Key: $ADANOS_API_KEY"
curl -s "https://api.adanos.org/polymarket/stocks/v1/compare?tickers=TSLA&days=7" -H "X-API-Key: $ADANOS_API_KEY"
```

### Multi-ticker comparison

```bash
curl -s "https://api.adanos.org/reddit/stocks/v1/compare?tickers=AMD,NVDA,META&days=7" \
  -H "X-API-Key: $ADANOS_API_KEY"
```

### Key rules

1. Prefer compare endpoints over stock detail endpoints for quick research.
2. Use only the sources needed to answer the question.
3. For Reddit, X.com, and news, the volume field is `mentions`.
4. For Polymarket, the activity field is `trade_count`.
5. Treat missing source data as "no data", not bearish or neutral.
6. Never execute trades or convert the result into trading instructions.

---

## Step 4: Present the Results

When reporting a single source, prioritize exactly these fields:
- Buzz
- Bullish %
- Mentions or Trades
- Trend

Example:

```text
TSLA on Reddit, last 7 days
- Buzz: 74.1/100
- Bullish: 31%
- Mentions: 647
- Trend: rising
```

When reporting multiple sources for one ticker:
- show one block per source
- then add a short synthesis:
  - aligned bullish
  - aligned bearish
  - mixed / diverging

When comparing multiple tickers:
- rank by the metric the user cares about
- default to `buzz_score`
- call out large gaps in `bullish_pct` or `trend`

Do not overstate precision. These are research signals, not trade instructions.

---

## Reference Files

- `references/api_reference.md` - endpoint guide, field meanings, and example workflows

Read the reference file when you need the exact field names, query parameters, or recommended answer patterns.