Skip to content

AGENTS.md

Latest commit

 

History

History
96 lines (70 loc) · 2.6 KB

AGENTS.md

File metadata and controls

96 lines (70 loc) · 2.6 KB

Agent Guidelines

This document provides guidance for AI agents using the mcp-multiutil MCP server.

Tool Descriptions

key-val

Store and retrieve simple key-value pairs. Useful for maintaining state, caching results, or storing configuration.

Common Use Cases:

  • Storing user preferences
  • Caching computed results
  • Maintaining session state

vector-store

Store vector embeddings with associated metadata. Enables semantic search and similarity matching.

Common Use Cases:

  • Semantic search over documents
  • Finding similar items
  • Building recommendation systems

rag

Retrieval-augmented generation using stored documents and their vector embeddings.

Common Use Cases:

  • Question answering over a knowledge base
  • Contextual information retrieval
  • Document-based reasoning

embedding

Generate and store text embeddings for semantic operations.

Common Use Cases:

  • Creating searchable vector representations
  • Preparing data for RAG
  • Similarity computations

txt-to-json

Convert unstructured text into structured JSON with automatic schema inference.

Common Use Cases:

  • Parsing natural language into structured data
  • Extracting entities and relationships
  • Data normalization

txt-to-toon

Transform text into ASCII art or cartoon representations.

Common Use Cases:

  • Creating visual text representations
  • Generating ASCII diagrams
  • Fun text transformations

store-md

Store markdown documents with searchable metadata.

Common Use Cases:

  • Building a markdown-based knowledge base
  • Storing documentation
  • Managing notes and articles

token-count

Count the number of tokens in a given text.

Common Use Cases:

  • Estimating API costs
  • Checking context window limits
  • Text analysis

tokenize

Split text into individual tokens.

Common Use Cases:

  • Text preprocessing
  • Analyzing token distribution
  • Building custom NLP pipelines

store-context

Store conversation context for later retrieval and reference.

Common Use Cases:

  • Maintaining conversation history
  • Building conversational memory
  • Context-aware responses

Best Practices

  1. Data Persistence: All data is stored persistently in libSQL. Clean up old data when no longer needed.
  2. Error Handling: Tools will return descriptive errors. Always check results before proceeding.
  3. Performance: The database runs in-memory for fast operations, but large datasets may impact performance.
  4. Semantic Operations: Use embeddings and vectors for semantic search rather than exact string matching.

Examples

(Coming soon - specific examples will be added as tools are implemented)