Add HuggingFace sentiment analysis skill category

[reggiezo]

HuggingFace Skill Category

This skill category enables natural language understanding tools via HuggingFace Transformers.

Included Skills

🧠 Sentiment Analysis

Description:
Analyzes the sentiment of input text (e.g., "I love this!") and returns a label (POSITIVE, NEGATIVE, etc.) along with a confidence score.

Configuration Example

skills:
  huggingface:
    states:
      sentiment_analysis: public

[hyacinthus]

Could you please list some use cases for this skill? Since all LLMs can inherently understand this natural language information, it seems we don't need a skill to assist.

[taiyangc]

To expedite the review process, please join https://discord.com/invite/crestal, open a support ticket to apply for an intentkit dev role. We have a discussion channel there for you to join up with the rest of the developers.

[bluntbrain]

Hi @reggiezo
thanks for the contribution!

As @hyacinthus said, could you share a few use cases for this skill? Since LLMs already understand this kind of input pretty well, just trying to get a sense of what this adds.

[reggiezo]

Hi @hyacinthus @bluntbrain @taiyangc thank you for your comments

I detailed few of the use cases of the huggingface Sentiment skill.

It is important to note that while general LLMs can infer sentiment they lack specialization. HuggingFace models, fine-tuned on extensive labeled datasets, excel in sentiment classification, offering accuracy, consistency, and reliability for critical tasks like moderation, market analysis, and customer satisfaction. These models provide structured outputs (e.g., POSITIVE, NEGATIVE, NEUTRAL) with confidence scores, essential for repeatable results in dashboards and automated pipelines.

Use Case 1: Dashboards: Visualizing Sentiment Trends
A product team tracks customer sentiment across reviews, social media, or support tickets:

Input: Daily user-generated content (e.g., app reviews, tweets).
Processing: Sentiment Analysis skill labels each text.
Output: Structured labels (POSITIVE, NEGATIVE, etc.) and confidence scores are stored in a database.
Dashboard: Visualizes sentiment trends over time, by region, or product line.
Unlike general LLMs, HuggingFace ensures deterministic, actionable insights.

Use Case 2: Automated Pipelines - Triggering Actions
A customer service workflow triages support tickets:

Input: Ticket message: "This is ridiculous, I’ve waited two weeks!"
Output: NEGATIVE with 0.97 confidence.
Pipeline Reaction: Flags as urgent, escalates to an agent, and triggers an apology email.
Similarly, sales teams can leverage real-time positive sentiment spikes for targeted campaigns.

Use Case 3: Educational Content Surveys
For Openledger’s Edu model, where users rate article content, this skill could assess whether the content has a "Positive" or "Negative" impact, enhancing user feedback.

In summary, while general LLMs produce probabilistic text, HuggingFace models deliver structured, deterministic outputs, making them ideal for reliable, actionable applications.

[hyacinthus]

where is _analyzer ?