# Sentiment News

> Computes ticker/sector news sentiment for DSE stocks and separates rumour-driven items from fundamentals-driven news, since the Dhaka Stock Exchange is rumour-heavy. Use when the user asks about news sentiment, "what's the news saying about GP", headline tone, rumour vs real news, or wants a sentiment score to feed signal-synthesizer.

- Skill: `kuntal-r-d/sentiment-news-2` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add kuntal-r-d/sentiment-news-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/kuntal-r-d/sentiment-news-2/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- License: Apache-2.0
- Author: kuntal-r-d (https://skillmd.com/u/kuntal-r-d)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/kuntal-r-d/sentiment-news-2

---

# Sentiment & News

## When to use

Activate when a user wants the news read on a DSE stock or sector: "what's the
sentiment on GP?", "is the news good or bad?", "is this just a rumour?", "score the
headlines". This is the Sentiment Agent (PRD-002 REQ-019). Its score is designed to
feed `signal-synthesizer` alongside the technical and fundamental legs.

## What it does

Reads the shared contract's `news` array and produces an aggregate sentiment score in
`[-1, +1]`. Because the DSE is rumour-heavy (PRD-001 REQ-023), it explicitly
**separates rumour-driven from fundamentals-driven items**:

- **Lexicon scoring** — each headline is scored from positive vs negative keyword
  counts (stdlib only; no NLP libraries).
- **Rumour detection** — an item is a `RUMOUR` if its `source` is social/forum/
  unconfirmed/rumour, its category is rumour-like, or its headline mentions
  rumour/unconfirmed/speculation. Rumours are down-weighted to **x0.3** and, when
  they are the *only* material signal, the score is capped to `[-0.4, +0.4]` so a
  rumour alone can never drive a strong signal.
- **Fundamental classification** — items in `earnings`, `regulatory`,
  `corporate_action`, or `macro` carry full weight.

Output is a Thinking Card (see suite README).

## How to run

```bash
python3 scripts/sentiment.py --input data.json --pretty
cat data.json | python3 scripts/sentiment.py
```

Reads `news` (required) and optional `mode`/`ticker`/`as_of`. Returns `score`
(-1..+1, fundamental-weighted), `confidence`, `rating` (positive/neutral/negative),
`key_metrics` (item/fundamental/rumour counts, average fundamental sentiment),
per-item `reasoning`, and `flags`.

Try it now with the shared fixture:

```bash
python3 scripts/sentiment.py --input ../_fixtures/sample_input.json --pretty
```

## Interpreting output

- `rating`: `positive` (score >=0.2), `negative` (<=-0.2), else `neutral`.
- `flags`: `rumour_dominated` (no fundamental items, score capped) and
  `thin_coverage` (<2 items) both lower confidence — never silent drops.
- Compare `avg_fundamental_sentiment` against the headline `score` to see how much of
  the read is fundamentals vs noise.

## Notes

The lexicon, rumour-vs-news separation rationale, and DSE rumour context are
documented in [references/SENTIMENT.md](references/SENTIMENT.md). Output is
educational analysis only, never financial advice.