Module llm_cli 

LLM CLI Ecosystem Integration

Integrates Simon Willison’s LLM CLI tool for multi-model orchestration, embeddings, clustering, and RAG pipelines.

Usage

use reasonkit::thinktool::llm_cli::{LlmCliClient, EmbeddingConfig, ClusterConfig};

let client = LlmCliClient::new()?;

// Execute a prompt
let response = client.prompt("Analyze this code", Some("claude-sonnet-4")).await?;

// Generate embeddings
let embeddings = client.embed("text to embed", None).await?;

// Cluster documents
let clusters = client.cluster("documents", 5, None).await?;

Security

All user-provided inputs are validated and sanitized before being passed to shell commands. This includes:

• Model names (alphanumeric, hyphens, underscores, slashes, colons, dots)
• Collection names (alphanumeric, hyphens, underscores)
• Database paths (validated as safe filesystem paths)
• Template names (alphanumeric, hyphens, underscores)

Structs

ClusterAssignment

Individual cluster assignment

ClusterConfig

Configuration for clustering operations

ClusterResult

Result from clustering operation

EmbeddingConfig

Configuration for embedding operations

EmbeddingResult

Result from embedding operation

LlmCliClient

LLM CLI Client for multi-model orchestration

LlmCliConfig

Configuration for the LLM CLI client

PromptResult

Result from an LLM prompt

RagConfig

Configuration for RAG operations

SimilarityMatch

Individual similarity match

SimilarityResult

Result from similarity search