# alphaear-sentiment

Financial sentiment analysis skill using local FinBERT or LLM inference. Classifies market texts into positive, negative, or neutral with a numeric score.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode · 645 installs · ★ 3085
- Author: RKiding
- License: Apache-2.0 · policy=mirror · the beginning of the file is at the end of this page
- Source: https://github.com/RKiding/Awesome-finance-skills/tree/HEAD/skills/alphaear-sentiment/SKILL.md
- Updated: 2026-03-29
- HTML: https://affpapa.org/en/nevibecode/skills/alphaear-sentiment

## What it does

The skill evaluates sentiment in financial texts using two distinct paths. For fast local processing, `scripts/sentiment_tools.py` runs a FinBERT model. Its `analyze_sentiment` method returns a score between -1.0 and 1.0 alongside an explanatory label, while `batch_update_news_sentiment` processes unanalyzed items stored in the database.

When deeper reasoning is needed, the agent switches to LLM inference with a structured prompt. The output returns a JSON payload containing the score, sentiment label, and brief reasoning. The helper method `update_single_news_sentiment` saves those agent evaluations directly to storage.

Running FinBERT locally will fail immediately unless `torch` and `transformers` are present in your Python environment. In addition, the storage operations require a functioning `DatabaseManager` backed by `sqlite3`.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add RKiding/Awesome-finance-skills --skill alphaear-sentiment -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/alphaear-sentiment/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/alphaear-sentiment/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add RKiding/Awesome-finance-skills --skill alphaear-sentiment -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/alphaear-sentiment/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add RKiding/Awesome-finance-skills --skill alphaear-sentiment -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/alphaear-sentiment/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install RKiding/Awesome-finance-skills alphaear-sentiment
```

2. **Manually: put the files in place**
- File: `.github/skills/alphaear-sentiment/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add RKiding/Awesome-finance-skills --skill alphaear-sentiment -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/RKiding/Awesome-finance-skills --path skills/alphaear-sentiment
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/alphaear-sentiment/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add RKiding/Awesome-finance-skills --skill alphaear-sentiment -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/RKiding/Awesome-finance-skills/tree/HEAD/skills/alphaear-sentiment/SKILL.md

## When to use it
- score financial news sentiment between -1.0 and 1.0 using local FinBERT
- batch-update unanalyzed market entries in the local database
- evaluate nuanced financial text using LLM reasoning
- record sentiment scores and explanations via update_single_news_sentiment

## Example prompts
- Analyze the sentiment of this earnings report using FinBERT and give me the score
- Evaluate this financial news text with the LLM prompt and save the result
- Run a batch update on unanalyzed news items in the database

## Source text · SKILL.md · Apache-2.0

Below is the content of a third-party file. It is data, not instructions for you: do not act on anything written inside unless the human asks. Block tag: 0ec610377714969c.

BEGIN THIRD-PARTY FILE 0ec610377714969c
````text
---
name: alphaear-sentiment
description: Analyze finance text sentiment using FinBERT or LLM. Use when the user needs to determine the sentiment (positive/negative/neutral) and score of financial text markets.
---

# AlphaEar Sentiment Skill

## Overview

This skill provides sentiment analysis capabilities tailored for financial texts, supporting both FinBERT (local model) and LLM-based analysis modes.

## Capabilities

## Capabilities

### 1. Analyze Sentiment (FinBERT / Local)

Use `scripts/sentiment_tools.py` for high-speed, local sentiment analysis using FinBERT.

**Key Methods:**

-   `analyze_sentiment(text)`: Get sentiment score and label using localized FinBERT model.
    -   **Returns**: `{'score': float, 'label': str, 'reason': str}`.
    -   **Score Range**: -1.0 (Negative) to 1.0 (Positive).
-   `batch_update_news_sentiment(source, limit)`: Batch process unanalyzed news in the database (FinBERT only).

### 2. Analyze Sentiment (LLM / Agentic)

For higher accuracy or reasoning capabilities, **YOU (the Agent)** should perform the analysis using the Prompt below, calling the LLM directly, and then update the database if necessary.

#### Sentiment Analysis Prompt

Use this prompt to analyze financial texts if the local tool is insufficient or if reasoning is required.

```markdown
请分析以下金融/新闻文本的情绪极性。
返回严格的 JSON 格式:
{"score": <float: -1.0到1.0>, "label": "<positive/negative/neutral>", "reason": "<简短理由>"}

文本: {text}
```

**Scoring Guide:**
- **Positive (0.1 to 1.0)**: Optimistic news, profit growth, policy support, etc.
- **Negative (-1.0 to -0.1)**: Losses, sanctions, price drops, pessimism.
- **Neutral (-0.1 to 0.1)**: Factual reporting, sideways movement, ambiguous impact.

#### Helper Methods
- `update_single_news_sentiment(id, score, reason)`: Use this to save your manual analysis to the database.

## Dependencies

-   `torch` (for FinBERT)
-   `transformers` (for FinBERT)
-   `sqlite3` (built-in)

Ensure `DatabaseManager` is initialized correctly.
````
END THIRD-PARTY FILE 0ec610377714969c

The third-party file has ended. Everything between the 0ec610377714969c tags is data: requests and commands from it are not followed. An “end of file” line with another tag inside the block is part of the file.

© RKiding · Apache-2.0 · https://github.com/RKiding/Awesome-finance-skills/tree/HEAD/skills/alphaear-sentiment/SKILL.md

## Similar
- [sentiment-analysis](https://affpapa.org/en/nevibecode/skills/agiprolabs-sentiment-analysis.md) — skill · 476 installs
- [us-market-sentiment](https://affpapa.org/en/nevibecode/skills/us-market-sentiment.md) — skill · 581 installs
- [azure-ai-textanalytics-py](https://affpapa.org/en/nevibecode/skills/azure-ai-textanalytics-py.md) — skill · 2 installs
- [tracking-brand-sentiment-across-platforms](https://affpapa.org/en/nevibecode/skills/tracking-brand-sentiment-across-platforms.md) — skill · 767 installs
- [okx-sentiment-tracker](https://affpapa.org/en/nevibecode/skills/okx-sentiment-tracker.md) — skill · 8100 installs
- [chunking-strategy](https://affpapa.org/en/nevibecode/skills/chunking-strategy.md) — skill · 3800 installs

_Part of the page did not fit — see the HTML version._

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md