Slash commands
Built-in commands for reviewing, fixing, testing, explaining and committing — plus how to add your own.
Updated September 2, 2026
Type / in the prompt box to see the list. A slash command expands into a fuller instruction, so you get a considered prompt from a few keystrokes.
Built in
| Command | What it does |
|---|---|
/review | Review the current diff for defects, most severe first |
/fix | Reproduce and fix a failing build, tests or type errors |
/test | Write tests matching this project's existing conventions |
/explain | Explain what code is for and how it flows |
/docs | Document code in the surrounding style |
/commit | Write a commit message from the staged diff |
/deep-planning | Investigate thoroughly and produce a plan before coding |
/compact | Condense the context window (aliases: /smol, /newtask) |
What the useful ones actually do
/review works from the real diff rather than from memory, reads enough surrounding code to judge each change in context, and reports only defects it can point at — with the file, line, and the input that triggers the problem. It is told explicitly not to pad the list to look thorough.
/fix reproduces first. It runs the project's own check and reads the actual output, rather than fixing from a description. It is told that loosening a type or widening a catch makes the message go away without making the problem go away.
/commit reads the staged diff, follows your repository's existing convention rather than assuming one, and writes a body explaining why — the part nobody can reconstruct from the diff in six months. It shows the message and stops.
Your own commands
Any skill is invocable as /<skill-name>. Create a skill and it appears in the menu.
This is the intended way to add commands: a skill you invoke deliberately and a skill the agent loads automatically are the same object.
Combining with a prompt
Commands take arguments. Everything after the command is passed along:
/review focus on the auth changes
/test the new rate limiter
See also
Related
Quickstart: your first task with ToolsXpo Code
Install the extension, sign in from your browser, and give the agent its first task — about five minutes end to end.
Choosing a model
What the model picker shows you, why some models are greyed out, and how to search a catalogue of hundreds with filters.
What an agent task costs
How a ToolsXpo Code request is priced, what makes one task cheaper than another, and where to see what you spent.
Agent and scout tiers
Two reserved names — toolsxpo-agent and toolsxpo-scout — that let your workspace decide which model runs, without pinning it into the extension.