Update project documentation

.github/workflows/ci.yml

@@ -20,10 +20,8 @@ jobs:
         uses: astral-sh/setup-uv@v6
       - name: Sync dependencies
         run: uv sync --group dev
-      - name: Lint (Ruff)
+      - name: Lint
         run: uv run lint
-      - name: Black
-        run: uv run black --check .
       - name: Typecheck (mypy)
         run: uv run typecheck
 
@@ -67,9 +65,25 @@ jobs:
       - name: Evaluations (Mock Mode)
         run: uv run python -m intent_kit.evals.run_all_evals --quiet --mock
 
+  security:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: '3.11'
+      - name: Install uv
+        uses: astral-sh/setup-uv@v6
+      - name: Sync dependencies
+        run: uv sync --group dev
+      - name: Install pip-audit
+        run: uv add --group dev pip-audit
+      - name: Security audit
+        run: uv run pip-audit --local
+
   build:
     runs-on: ubuntu-latest
-    needs: [test, eval]
+    needs: [test, eval, security]
     steps:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v5

.pre-commit-config.yaml

@@ -46,6 +46,16 @@ repos:
         language: system
         files: .codecov.yml
         pass_filenames: false
+      - id: security
+        name: Security audit
+        entry: uv run security
+        language: system
+        pass_filenames: false
+      - id: auto-amend
+        name: Auto-amend commit with reformatted files
+        entry: uv run auto-amend
+        language: system
+        pass_filenames: false
   - repo: https://github.com/pre-commit/pre-commit-hooks
     rev: v4.5.0
     hooks:

README.md

@@ -260,6 +260,11 @@ This means you can deploy with confidence, knowing your AI workflows work reliab
 - Catch regressions automatically
 - Validate reliability before deployment
 
+### **Security**
+- Automated security audits with pip-audit
+- Vulnerability scanning in CI/CD pipeline
+- Dependency security monitoring
+
 ---
 
 ## Common Use Cases
@@ -340,8 +345,9 @@ uv run pre-commit install
 ```bash
 uv run pytest          # Run tests
 uv run lint            # Lint code
-uv run black --check . # Format check
+uv run black --fix .   # Format and fix code
 uv run typecheck       # Type checking
+uv run security        # Security audit
 uv build               # Build package
 ```
 

docs/concepts/nodes-and-actions.md

@@ -44,6 +44,82 @@ weather_action = action(
 - **action_func** - Function to execute
 - **param_schema** - Parameter type definitions
 
+#### Argument Extraction
+
+Actions automatically extract parameters from user input using the argument extraction system:
+
+- **RuleBasedArgumentExtractor** - Uses pattern matching and rules for fast extraction
+- **LLMArgumentExtractor** - Uses LLM for intelligent parameter extraction
+- **Automatic Selection** - Intent Kit chooses the best extractor based on your configuration
+
+```python
+from intent_kit import action
+
+# Rule-based extraction (fast, deterministic)
+greet_action = action(
+    name="greet",
+    description="Greet the user",
+    action_func=lambda name: f"Hello {name}!",
+    param_schema={"name": str},
+    argument_extractor="rule"  # Use rule-based extraction
+)
+
+# LLM-based extraction (intelligent, flexible)
+weather_action = action(
+    name="weather",
+    description="Get weather information",
+    action_func=lambda city: f"Weather in {city} is sunny",
+    param_schema={"city": str},
+    argument_extractor="llm"  # Use LLM extraction
+)
+```
+
+#### Error Handling Strategies
+
+Actions support pluggable error handling strategies for robust execution:
+
+```python
+from intent_kit import action
+
+# Retry on failure
+retry_action = action(
+    name="retry_example",
+    description="Example with retry strategy",
+    action_func=lambda x: x / 0,  # Will fail
+    param_schema={"x": float},
+    remediation_strategy="retry_on_fail",
+    remediation_config={
+        "max_attempts": 3,
+        "base_delay": 1.0
+    }
+)
+
+# Fallback to another action
+fallback_action = action(
+    name="fallback_example",
+    description="Example with fallback strategy",
+    action_func=lambda x: x / 0,  # Will fail
+    param_schema={"x": float},
+    remediation_strategy="fallback_to_another_node",
+    remediation_config={
+        "fallback_name": "safe_calculation"
+    }
+)
+
+# Self-reflection for parameter correction
+reflect_action = action(
+    name="reflect_example",
+    description="Example with self-reflection",
+    action_func=lambda name: f"Hello {name}!",
+    param_schema={"name": str},
+    remediation_strategy="self_reflect",
+    remediation_config={
+        "max_reflections": 2,
+        "llm_config": {"provider": "openai", "model": "gpt-3.5-turbo"}
+    }
+)
+```
+
 ### Classifier Nodes
 
 Classifier nodes route input to appropriate child nodes based on classification logic.

docs/development/documentation-management.md

@@ -96,7 +96,7 @@ The documentation has been updated to use consistent terminology:
 
 - **Actions** instead of "handlers" - Functions that execute and produce outputs
 - **Classifiers** - Nodes that route input to appropriate actions
-- **Splitters** - Nodes that handle multiple nodes in single input
+- **Classifiers** - Nodes that route input to appropriate actions
 
 ## Navigation Structure
 

docs/development/index.md

@@ -8,5 +8,6 @@ Welcome to the Development section of the Intent Kit documentation. Here you'll
 - [Testing](testing.md): Unit tests and integration testing.
 - [Evaluation](evaluation.md): Performance evaluation and benchmarking.
 - [Debugging](debugging.md): Debugging tools and techniques.
+- [Performance Monitoring](performance-monitoring.md): Performance tracking and reporting.
 
 For additional information, see the [project README on GitHub](https://github.com/Stephen-Collins-tech/intent-kit#readme) or explore other sections of the documentation.