Contextual Anchor Design

knowledge skill

Identifies and extracts key contextual anchors from unstructured text to label and structure knowledge for retrieval or downstream processing. Produces a structured list of anchors with their context and relevance scores.

Worked examples

• {"input": {"text": "The patient presented with acute chest pain radiating to the left arm, diaphoresis, and nausea. Initial vitals: BP 160/90, HR 110. ECG showed ST elevation in leads V1-V4.", "domain": "medical", "max_anchors": 5}, "output": {"anchors": [{"anchor": "acute chest pain radiating to left arm", "context": "symptom presentation", "relevance": 0.98}, {"anchor": "ST elevation in leads V1-V4", "context": "ECG finding", "relevance": 0.95}, {"anchor": "BP 160/90", "context": "initial vitals", "relevance": 0.85}], "processing_time_ms": 42}}
• {"input": {"text": "The model achieved 92% accuracy on the ImageNet validation set with a ResNet-50 backbone using SGD optimizer and learning rate 0.01.", "domain": "machine learning", "max_anchors": 3}, "output": {"anchors": [{"anchor": "92% accuracy on ImageNet validation set", "context": "model performance", "relevance": 0.99}, {"anchor": "ResNet-50 backbone", "context": "model architecture", "relevance": 0.96}, {"anchor": "SGD optimizer with learning rate 0.01", "context": "training configuration", "relevance": 0.92}], "processing_time_ms": 38}}

Input

• text: str
• domain: str
• max_anchors: int

Output

• anchors: list[dict]
• processing_time_ms: int

Details

• Skill type: knowledge skill
• Safety level: safe_public_research
• Version: 1.0.0