Slash Commands

Slash commands are the fastest way to discover and trigger Dawn commands from chat. Some commands are built-in product actions; Prompt Commands are reusable workspace prompt workflows that start a normal agent run.

Two Kinds Of Slash Commands

Product commands

Built-in commands such as /help, /whats_new, and /link perform predictable Dawn product actions. They are not open-ended prompts.

Prompt Commands

Prompt Commands such as /prototype, /pr-review, /ticket-plan, and /reflect are reusable stored prompt workflows. They use the selected agent and current workspace context.

For creating, copying, editing, and restoring Prompt Commands, see Prompt Commands.

Why They Matter

They make common Dawn actions and workspace commands discoverable while you are typing.
They keep deterministic product actions separate from prompt-backed agent workflows.
They let users confirm exact command usage with /help instead of memorizing syntax.

Built-In Product Commands

/help shows the command list and usage for the current surface.
/help <command> shows the usage for one specific command.
/whats_new returns recent Dawn product updates from the public Dawn changelog.
/link [email protected] starts identity linking in supported private external chats.
/link status checks whether the current external identity is already linked where that command is supported.

Prompt Commands In The Same Composer

Supported Dawn chat surfaces can also show workspace Prompt Commands in the slash-command picker. These commands start a normal agent run with stored instructions attached. See the Command Reference for every shipped command page.

Use /prototype <brief> for grounded HTML artifacts and lightweight prototypes.
Use /pr-review <PR URL, branch, or focus> for correctness-focused code review workflows.
Use /reflect <focus> to audit recent workspace context and suggest memory, skill, command, or workflow improvements.
Use /ticket-plan <ticket URL, key, or brief> to turn work items into implementation plans.
Use /release-notes <range, milestone, PRs, or tickets> for user-facing release-note drafts.

The exact list depends on the current workspace. Global Prompt Commands can be copied into workspace commands and customized from Dawn Web.

Where To Use Them

In Dawn’s first-party chat surfaces, typing / opens command suggestions directly in the composer.
/help and /whats_new are normal starting points in Dawn Web and other first-party chat surfaces that expose slash commands.
Prompt Commands are available wherever Dawn exposes the workspace command catalog for that chat surface.
/link matters most in supported private external chats such as Slack DMs, Telegram private chats, Teams personal chats, and Discord DMs because identity linking starts from those private surfaces.

How To Use Them Well

Type / at the start of the draft so Dawn can show command suggestions.
Use /help first if you want to confirm which commands the current surface exposes.
Use /help <command> when you want the exact usage for one command.
Use /whats_new when you want product updates about Dawn itself rather than workspace-specific operational notes.
Use /link only from the private external chat identity you actually want to connect to your Dawn account.
Use Prompt Commands when you want a normal model answer shaped by reusable workspace instructions.

What Users Often Misunderstand

/whats_new is about Dawn product updates, not internal workspace announcements.
/link is about identity linking, not agent linking or workspace invitations.
Prompt Commands do not grant extra tools, skills, memory, or connected-source access. The selected agent still controls available context and permissions.
If a command is unavailable on a surface, that usually means the surface does not expose that command, not that Dawn failed to understand the slash input.

Troubleshooting

Slash commands do not appear: type / at the beginning of the draft and confirm you are in a supported surface.
/whats_new returns nothing useful: check whether the public Dawn changelog has recent published entries.
/link does not work where you expected: use a supported private external chat rather than a shared team channel.
A Prompt Command does not appear: confirm it exists in the workspace, is enabled, and is exposed by the current chat surface.
You expected a model answer but got an immediate command response: you ran a built-in product command, not a Prompt Command.

Completion Checklist

Slash-command suggestions appear where expected.
/help returns current command usage.
/whats_new returns recent Dawn product updates.
/link is only used from supported private external chats.
Users understand the difference between product slash commands and prompt-backed workspace commands.