Crate llm_coding_tools_rig

Expand description

§llm-coding-tools-rig

Lightweight, high-performance Rig framework Tool implementations for coding tools.

§Features

File operations - Read, write, edit, glob, grep with two access modes:
- absolute::* - Unrestricted filesystem access
- allowed::* - Sandboxed to configured directories
Shell execution - Cross-platform command execution with timeout
Web fetching - URL content retrieval with format conversion
Todo management - Shared-state todo list tracking
Context strings - LLM guidance text for tool usage (re-exported from core)

§Installation

Add to your Cargo.toml:

[dependencies]
llm-coding-tools-rig = "0.1"

§Quick Start

Minimal runnable agent (requires OPENAI_API_KEY):

use llm_coding_tools_rig::absolute::{GlobTool, GrepTool, ReadTool};
use llm_coding_tools_rig::{BashTool, SystemPromptBuilder, TodoTools};
use rig::providers::openai;
use rig::client::{ProviderClient, CompletionClient};
use rig::completion::Prompt;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let todos = TodoTools::new();
    let mut pb = SystemPromptBuilder::new();

    // Build agent with system prompt tracking
    let client = openai::Client::from_env();
    let agent = client
        .agent("gpt-4o")
        .tool(pb.track(ReadTool::<true>::new()))
        .tool(pb.track(GlobTool::new()))
        .tool(pb.track(GrepTool::<true>::new()))
        .tool(pb.track(BashTool::new()))
        .tool(pb.track(todos.read))
        .tool(pb.track(todos.write))
        .preamble(&pb.build())  // Build system prompt after tracking tools
        .build();

    let response = agent
        .prompt("Search for TODO comments in src/")
        .await?;
    println!("{response}");

    Ok(())
}

Example preamble output (truncated):

# Environment

Working directory: /home/user/project

# Tool Usage Guidelines

## `Read` Tool
Reads files from disk.
## `Bash` Tool
Executes shell commands.

§Usage

File tools come in absolute::* (unrestricted) and allowed::* (sandboxed) variants:

use llm_coding_tools_rig::absolute::{ReadTool, WriteTool};
use llm_coding_tools_rig::allowed::{ReadTool as AllowedReadTool, WriteTool as AllowedWriteTool};
use llm_coding_tools_rig::AllowedPathResolver;
use std::path::PathBuf;

let read = ReadTool::<true>::new();
let resolver = AllowedPathResolver::new([PathBuf::from("/home/user/project")]).unwrap();
let sandboxed_read: AllowedReadTool<true> = AllowedReadTool::new(resolver.clone());
let sandboxed_write = AllowedWriteTool::new(resolver);

Other tools: BashTool, WebFetchTool, TodoTools. Use SystemPromptBuilder to register tools and pass pb.build() to .preamble(). Set working_directory() so the environment section is populated. Context strings are re-exported in llm_coding_tools_rig::context (e.g., BASH, READ_ABSOLUTE).

§Examples

# Basic toolset setup with SystemPromptBuilder
cargo run --example basic -p llm-coding-tools-rig

# Sandboxed file access with allowed::* tools
cargo run --example sandboxed -p llm-coding-tools-rig

§License

Apache 2.0

Re-exports§

pub use absolute::EditArgs;
pub use absolute::EditTool;
pub use absolute::GlobArgs;
pub use absolute::GlobTool;
pub use absolute::GrepArgs;
pub use absolute::GrepTool;
pub use absolute::ReadArgs;
pub use absolute::ReadTool;
pub use absolute::WriteTool;
pub use absolute::WriteToolArgs;
pub use bash::BashArgs;
pub use bash::BashTool;
pub use todo::TodoReadArgs;
pub use todo::TodoReadTool;
pub use todo::TodoTools;
pub use todo::TodoWriteArgs;
pub use todo::TodoWriteTool;
pub use webfetch::WebFetchArgs;
pub use webfetch::WebFetchTool;

Modules§

absolute: Tools using llm_coding_tools_core::path::AbsolutePathResolver.
allowed: Tools using llm_coding_tools_core::path::AllowedPathResolver.
allowed_tools: Re-export allowed module tool types (namespaced to avoid conflicts)
bash: Shell command execution tool.
context: Tool context strings for LLM agents.
todo: Todo list management tools.
webfetch: Web content fetching tool.

Structs§

AbsolutePathResolver: Path resolver that requires absolute paths.
AllowedPathResolver: Path resolver that restricts access to allowed directories.
BashOutput: Result of shell command execution.
GlobOutput: Output from glob file matching.
GrepFileMatches: All matches within a single file.
GrepLineMatch: A single line match within a file.
GrepOutput: Output from grep search.
SystemPromptBuilder: Builder that tracks tools and generates formatted system prompts.
Todo: A single task item.
TodoState: Thread-safe shared state for todo list.
ToolOutput: Wrapper for tool output with truncation metadata.
WebFetchOutput: Result from URL fetch operation.

Enums§

EditError: Errors specific to edit operations.
TodoPriority: Task priority level.
TodoStatus: Task status.
ToolError: Unified error type for all tool operations.

Traits§

PathResolver: Strategy for resolving and validating file paths.
Substitute: Extension trait for placeholder substitution on system prompt strings.
ToolContext: Trait for tools that provide usage context for LLM system prompts.

Type Aliases§

ToolResult: Result type alias for tool operations.