# Taxonomy of AI Agent Skills | Category | Skill | UID | Description | |-----------------------------------------------|-----------------------------------------------|-------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | Natural Language Understanding | Natural Language Understanding | 101 | Describes the ability to understand the meaning of a given situation or event. | | Natural Language Generation | Natural Language Generation | 102 | Describes the ability to generate human-like text from structured data or other inputs. | | Information Retrieval and Synthesis | Information Retrieval and Synthesis | 103 | Capabilities for retrieving relevant information from various sources and synthesizing it into coherent, contextually appropriate responses. This includes searching, extracting, combining, and presenting information in a meaningful way. | | Creative Content Generation | Creative Content Generation | 104 | Capabilities for generating various forms of creative content, including narratives, poetry, and other creative writing forms. | | Language Translation and Multilingual Support | Language Translation and Multilingual Support | 105 | Capabilities for handling multiple languages, including translation and multilingual text processing. | | Personalisation and Adaptation | Personalisation and Adaptation | 106 | Capabilities for adapting and personalizing content based on user context and preferences. | | Analytical and Logical Reasoning | Analytical and Logical Reasoning | 107 | Capabilities for performing logical analysis, inference, and problem-solving tasks. | | Ethical and Safe Interaction | Ethical and Safe Interaction | 108 | Capabilities for ensuring ethical, unbiased, and safe content generation and interaction. | | Text Classification | Text Classification | 109 | Capabilities for classifying and categorizing text into predefined categories or labels. | | Feature Extraction | Feature Extraction | 110 | Capabilities for extracting and representing textual features as vectors for downstream tasks. | | Token Classification | Token Classification | 111 | Capabilities for classifying individual tokens or words within text. | | Contextual Comprehension | Contextual Comprehension | 10101 | Describes the ability to understand the context of a given situation or event. | | Semantic Understanding | Semantic Understanding | 10102 | Describes the ability to understand the meaning of a given situation or event. | | Entity Recognition | Entity Recognition | 10103 | Describes the ability to identify and classify entities in a given situation or event. | | Text Completion | Text Completion | 10201 | Continuing a given text prompt in a coherent and contextually appropriate manner to generate fluent and contextually relevant content. | | Text Summarization | Text Summarization | 10202 | Condensing longer texts into concise summaries while preserving essential information and maintaining coherence. | | Text Paraphrasing | Text Paraphrasing | 10203 | Rewriting text to express the same ideas using different words and structures while maintaining the original meaning. | | Dialogue Generation | Dialogue Generation | 10204 | Producing conversational responses that are contextually relevant and engaging within a dialogue context. | | Question Generation | Question Generation | 10205 | Automatically generating relevant and meaningful questions from a given text or context. | | Text Style Transfer | Text Style Transfer | 10206 | Rewriting text to match the style of a given reference text while preserving the original content. | | Story Generation | Story Generation | 10207 | Generating a piece of text given a description or a first sentence to complete. | | Fact Extraction | Fact Extraction | 10301 | Capability to identify and extract factual information from text documents or knowledge bases, including entities, relationships, and key data points. | | Question Answering | Question Answering | 10302 | System capability to understand questions and provide accurate, relevant answers by analyzing available information sources. | | Knowledge Synthesis | Knowledge Synthesis | 10303 | Capability to aggregate and combine information from multiple sources, creating comprehensive and coherent responses while maintaining context and relevance. | | Sentence Similarity | Sentence Similarity | 10304 | Capability to analyze and determine the semantic similarity between sentences, supporting tasks like search, matching, and content comparison. | | Document and Passage Retrieval | Document and Passage Retrieval | 10305 | Capability to identify and retrieve relevant documents or text passages based on specific criteria or queries from a larger collection of texts. | | Search | Search | 10306 | Capability to perform efficient and accurate searches within large textual databases based on various criteria, including keywords, semantic meaning, or complex queries. | | Storytelling | Storytelling | 10401 | Creating narratives, stories, or fictional content with creativity and coherence. | | Poetry and Creative Writing | Poetry and Creative Writing | 10402 | Composing poems, prose, or other forms of creative literature. | | Translation | Translation | 10501 | Converting text from one language to another while maintaining meaning and context. | | Multilingual Understanding | Multilingual Understanding | 10502 | Recognising and processing text in multiple languages. | | User Adaptation | User Adaptation | 10601 | Tailoring responses based on user preferences, history, or context. | | Tone and Style Adjustment | Tone and Style Adjustment | 10602 | Modifying the tone or style of generated text to suit specific audiences or purposes. | | Inference and Deduction | Inference and Deduction | 10701 | Making logical inferences based on provided information. | | Problem Solving | Problem Solving | 10702 | Assisting with solving problems by generating potential solutions or strategies. | | Fact and Claim Verification | Fact and Claim Verification | 10703 | Verifying facts and claims given a reference text. | | Bias Mitigation | Bias Mitigation | 10801 | Reducing or eliminating biased language and ensuring fair and unbiased output. | | Content Moderation | Content Moderation | 10802 | Avoiding the generation of harmful, inappropriate, or sensitive content. | | Topic Labelling and Tagging | Topic Labelling and Tagging | 10901 | Classifying a text as belonging to one of several topics, which can be used to tag a text. | | Sentiment Analysis | Sentiment Analysis | 10902 | Classify the sentiment of a text, e.g., a positive movie review. | | Natural Language Inference | Natural Language Inference | 10903 | Classifying the relation between two texts, e.g., as a contradiction, entailment, etc. | | Model Feature Extraction | Model Feature Extraction | 11001 | Representing parts of text with vectors to be used as input to other tasks. | | Named Entity Recognition | Named Entity Recognition | 11101 | Task to recognise names as entities, e.g., people, locations, buildings, etc. | | Part-of-Speech Tagging | Part-of-Speech Tagging | 11102 | Tagging each part of a sentence as nouns, adjectives, verbs, etc. |