Without the right setup, coding agents guess at conventions, ignore boundaries, and produce inconsistent results across tools and repos. The Joggr Toolkit gives every agent the same foundation, regardless of which provider your team uses. You define your workspace once. Joggr generates the right files, wires the right context, and enforces the right boundaries for every provider in your stack.Documentation Index
Fetch the complete documentation index at: https://docs.joggr.ai/llms.txt
Use this file to discover all available pages before exploring further.
Context
Joggr organizes context into three layers — Repo Context, Org Context, and External Context — each scoped to where the information lives and what it’s for.jog init generates your AI instructions and rules. jog docs generates your internal docs. jog context connects org and external sources. jog drift keeps your repo context current.
Learn more about Context
Sandbox
Running multiple agents in the same repo means conflicts, overwritten files, and agents stepping on each other. Running one agent at a time means waiting. Sandboxes give each agent its own isolated environment — local worktree, container, or cloud VM — so agents can work in parallel without interfering with your working directory or each other. Attach and detach as needed, or let them run in the background while you do something else.jog sandbox creates and manages your sandboxes.
Learn more about Sandbox
Workflows
Complex features don’t fit in a single prompt. Without structure, you end up project-managing the agent yourself — breaking work into phases, feeding context between steps, and rubber-stamping 800-line diffs at the end. Workflows is the structured execution engine built into your AI coding tool. It takes a feature from idea through discussion, research, planning, execution, and verification — with human review gates at every phase. You review a plan before code gets written, and review each phase before it ships.jog gg drives the workflow. jog app gives you a local review UI for documents and diffs.
Learn more about Workflows
Guardrails
Agents make mistakes. They ignore standards, run destructive commands, and produce bad output. Without enforcement, you only find out after the damage is done. Guardrails catch problems before they land. Hooks block bad commits.jog drift flags stale docs before they mislead agents. jog eval confirms your agent skills work as expected.
Learn more about Guardrails
Compact
CLI output is designed for humans — progress bars, color codes, verbose logging. Agents don’t need any of it, and irrelevant tokens in the context window actively hurt output quality. Compact filters that noise so agents only see what matters: errors, failures, and results. The full command still runs — only what reaches the agent is filtered.jog compact wraps any command and cuts token usage by 60–90% depending on the operation.
Learn more about Compact