feat: add Claude Code plugin with skill and slash commands

Author: cardmagic

.claude-plugin/plugin.json

@@ -0,0 +1,12 @@
+{
+  "name": "classifier",
+  "description": "Text classification CLI using Bayesian, LSI, KNN, Logistic Regression, and TF-IDF algorithms",
+  "version": "1.0.0",
+  "author": {
+    "name": "Lucas Carlson",
+    "url": "https://github.com/cardmagic"
+  },
+  "repository": "https://github.com/cardmagic/classifier",
+  "license": "LGPL-2.1",
+  "keywords": ["machine-learning", "text-classification", "nlp", "bayesian", "lsi", "cli"]
+}

commands/classify.md

@@ -0,0 +1,32 @@
+# /classifier:classify
+
+Classify text using a pre-trained or custom model.
+
+## Usage
+
+```
+/classifier:classify <model> <text>
+```
+
+## Examples
+
+```
+/classifier:classify sms-spam-filter "Congratulations! You won $1000"
+/classifier:classify imdb-sentiment "This movie was fantastic"
+/classifier:classify emotion-detection "I feel so frustrated today"
+```
+
+## Instructions
+
+Run the classifier command with the specified model and text:
+
+```bash
+classifier -r "$model" "$text"
+```
+
+If no model is specified, list available models with `classifier models` and ask the user which one to use.
+
+Report the classification result clearly, e.g.:
+- "The text was classified as: **spam**"
+- "Sentiment: **positive**"
+- "Detected emotion: **anger**"

commands/models.md

@@ -0,0 +1,24 @@
+# /classifier:models
+
+List all available pre-trained classification models.
+
+## Usage
+
+```
+/classifier:models
+```
+
+## Instructions
+
+Run:
+
+```bash
+classifier models
+```
+
+Present the results in a formatted table showing:
+- Model name
+- Description/use case
+- Categories it can classify
+
+Suggest relevant models based on the user's context if known.

commands/train.md

@@ -0,0 +1,33 @@
+# /classifier:train
+
+Train a classifier with labeled examples.
+
+## Usage
+
+```
+/classifier:train <category> <text or file pattern>
+```
+
+## Examples
+
+```
+/classifier:train spam "Buy cheap viagra now"
+/classifier:train ham "Meeting scheduled for tomorrow"
+/classifier:train positive reviews/good/*.txt
+/classifier:train negative reviews/bad/*.txt
+```
+
+## Instructions
+
+Run the classifier train command:
+
+```bash
+classifier train "$category" "$text_or_pattern"
+```
+
+After training, inform the user they can:
+1. Add more training examples with additional `/classifier:train` commands
+2. Classify new text with `classifier "text to classify"`
+3. Save the model with `classifier save model-name.json`
+
+For best results, recommend balanced training data across all categories.

skills/classifier/SKILL.md

@@ -0,0 +1,67 @@
+# Text Classification with Classifier
+
+Use when: User asks to classify text, detect spam, analyze sentiment, detect emotions, or use pre-trained ML models.
+
+## Pre-trained Models
+
+Run `classifier models` to see all available models. Common ones:
+
+| Model | Command | Use Case |
+|-------|---------|----------|
+| `sms-spam-filter` | `classifier -r sms-spam-filter "text"` | Spam detection |
+| `imdb-sentiment` | `classifier -r imdb-sentiment "text"` | Sentiment analysis |
+| `emotion-detection` | `classifier -r emotion-detection "text"` | Emotion classification |
+
+## Quick Classification
+
+```bash
+# Classify with a pre-trained model
+classifier -r <model-name> "text to classify"
+
+# Example: detect spam
+classifier -r sms-spam-filter "You won a free iPhone! Click here now!"
+
+# Example: sentiment analysis
+classifier -r imdb-sentiment "This movie was absolutely terrible"
+
+# Example: emotion detection
+classifier -r emotion-detection "I am so happy today"
+```
+
+## Custom Training
+
+```bash
+# Train from text
+classifier train positive "Great product, love it"
+classifier train negative "Terrible quality, waste of money"
+
+# Train from files
+classifier train positive reviews/good/*.txt
+classifier train negative reviews/bad/*.txt
+
+# Classify after training
+classifier "This product exceeded my expectations"
+```
+
+## Model Management
+
+```bash
+# List all available models
+classifier models
+
+# Show model details
+classifier info <model-name>
+
+# Save trained model
+classifier save my-model.json
+
+# Load saved model
+classifier load my-model.json
+```
+
+## Best Practices
+
+1. For quick classification tasks, use pre-trained models first
+2. For custom domains, train with representative examples from each category
+3. Use `classifier models` to discover available pre-trained models
+4. Balance training data across categories for best results