Sentiment Analysis

Sentiment Analysis

Systematic sentiment scoring that quantifies market psychology to identify extremes, confirm setups, and flag contrarian opportunities.

Source Hierarchy

| Source Type | Weight | Examples | | --- | --- | --- | | Earnings calls / SEC filings | 5 | 10-K, 10-Q, transcripts | | Financial wire services | 4 | Reuters, Bloomberg, AP | | Analyst reports / ratings | 3 | Goldman, Morgan Stanley | | Financial news sites | 2 | CNBC, MarketWatch, Seeking Alpha | | Social media / forums | 1 | Twitter/X, Reddit, StockTwits |

Article Scoring (1-10)

| Score | Meaning | Examples | | --- | --- | --- | | 1-2 | Strongly bearish | Bankruptcy risk, SEC investigation, massive miss | | 3-4 | Moderately bearish | Guidance cut, downgrade, margin compression | | 5 | Neutral | Routine filings, expected results | | 6-7 | Moderately bullish | Earnings beat, upgrade, new product launch | | 8-9 | Strongly bullish | Massive beat + raised guidance, transformative deal | | 10 | Euphoric | "Best quarter ever" -- often a contrarian warning |

Temporal Decay

| News Age | Decay Weight | | --- | --- | | 3 days | 0.1 |

Composite Score

weighted_score = article_score x source_weight x decay_weight
Composite = sum(weighted_scores) / sum(source_weight x decay_weight)

| Composite | Classification | Signal | | --- | --- | --- | | 1.0 - 2.5 | Extreme fear | Contrarian buy zone (if at support) | | 2.5 - 4.0 | Bearish | Confirms bearish setups | | 4.0 - 6.0 | Neutral | No sentiment edge, rely on technicals | | 6.0 - 7.5 | Bullish | Confirms bullish setups | | 7.5 - 9.0 | Strong bullish | High conviction longs | | 9.0 - 10.0 | Extreme euphoria | Contrarian sell zone (if at resistance) |

Contrarian Signals

• Composite >9.0 at resistance = market top risk
• Composite <2.5 at support = market bottom opportunity
• Sentiment divergence: Price new highs + sentiment declining = distribution warning
• Sentiment divergence: Price new lows + sentiment improving = accumulation signal

Workflow

1. Gather News Data

get_financial_news(topic="AAPL earnings revenue guidance analyst", max_results=20)
get_financial_news(topic="AAPL analyst upgrade downgrade price target", max_results=10)

Search both bullish and bearish angles for the same asset.

2. Score Each Article

For each article: assign sentiment score (1-10), identify source type for weight, note publication time for decay.

3. Calculate Composite

Apply the formula. Report: composite score, classification, key drivers, contrarian signals if any, and confidence level (High = many sources agree, Medium = mixed, Low = few sources).

4. Identify Cycle Phase

Disbelief (negative sentiment, rising price) = accumulate. Euphoria (extreme positive, peak) = distribute. Panic/capitulation (despair, bottom) = watch for accumulation.

Key Rules

• NEVER score by gut feel -- follow the framework for every article systematically
• NEVER treat all sources equally -- a Reddit post is not a Bloomberg wire
• NEVER give equal weight to week-old news -- apply temporal decay
• NEVER use sentiment as primary entry signal -- it confirms technical setups
• NEVER only seek confirming news -- always search both bullish and bearish angles
• When high-credibility sources disagree, reduce conviction rather than picking a side
• Extreme sentiment can persist for weeks -- use as confirmation, not sole trigger

Related Skills

• earnings-trading -- Management tone scoring feeds sentiment
• economic-calendar-trading -- Pre-event sentiment gauges positioning

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: SKE-Labs
• Source: SKE-Labs/agent-trading-skills
• License: Apache-2.0

Install and usage instructions live in the source repository linked above.