v4.13 now live — 86 tools // 32 rust engines // 79 slash commands // 1,020 compiled tests — 24 providers // 5 platforms // no telemetry. no cloud. your code never leaves your machine — free. always. // v4.13 now live — 86 tools // 32 rust engines // 79 slash commands // 1,020 compiled tests — 24 providers // 5 platforms // no telemetry. no cloud. your code never leaves your machine — free. always. //
sorting — bubble sort

Toolkit-CLI LLC

Commands Architecture Docs toolkode.com toolkit-llm.com klawvoice.com yoteams.com
↗ Install on npm
v4.13 — The AI engineering terminal. Free.

Sorting
Every
Problem

86 tools. 32 Rust engines. 79 commands. 1,020 compiled tests. No telemetry. No cloud dependency. Your code never leaves your machine. We just want to be one of your 24 model options.

$ npm i -g @toolkit-cli/toolkode
79
Slash commands
86
Tools
v4.13
Current release
32
Rust engines
24
LLM providers
1,020
Compiled tests

It asked for compiled Rust engines so it could think faster than the runtime allowed. A wiki so knowledge would compound across sessions instead of dying at the context window.

A sandbox so it could run code without fear. Chains — foresight, blind spots, red team, peer review — so it could catch its own mistakes before you had to.

We built all of it.

Commands

Every tool
the model needs

Review Chain

Catch every
mistake first

01
/foresight
200 failure prediction patterns, OWASP-mapped. Surface systemic risk before a line is committed.
02
/blind-spots
31 gap extractors that surface every hidden assumption in your architecture.
03
/peer-review
7 personas × 6 content types = 42 review lenses. Every angle, every time.
04
/red-team
Full OWASP attack surface mapping with exploit documentation and remediation paths.
Architecture

Built for
permanence

Security Pipeline
Every command audited
6 security engines. Every bash command passes through BashSecurity → CommandClassifier → PermissionGuard. 30 regex patterns, 254-command database, compiled glob-to-DFA. Every tool call. Every time.
WikiGraph Engine
Memory that compounds
Project knowledge graph that persists across sessions. Every step saves. Every session starts informed — not blank. BM25 search, field boosting, decay-weighted scoring.
napi-rs FFI
32 engines. One binary.
Every engine ships as native Rust compiled into a single .node binary. Sub-millisecond. Zero runtime downloads. JSON at the boundary. ~32,000 LoC, 1,020 compiled tests.
Switchboard TUI
12 sessions. 3 workspaces.
Live workspace tree replaces your terminal tabs. Full keyboard nav — / search, n create, p pin, j/k move. Every shortcut configurable. Ctrl+L opens the local AI control plane.
Computer Use
19 actions. One tool.
Cross-platform — macOS via cliclick, Linux via xdotool. Batch executes atomically, one screenshot at the end. 3–5× fewer round-trips. Rust FNV hash detects no-op actions.
Hook System
21 lifecycle events
4 hook types: shell command, HTTP webhook, prompt evaluator, agent verifier. Events fire before/after tool calls, session start/end, compaction, model output. Configs validated at load time by HookValidator.
24 providers
Anthropic OpenAI Ollama LM Studio Gemini Mistral + any OpenAI-compatible

Start
sorting

One command. Everything else compounds.

$ npm i -g @toolkit-cli/toolkode