Attach files, folders, GitHub issues, GitHub PRs, and images to your Agent Chat prompts.

Attachments

Agent Chat lets you attach context to a message instead of pasting it inline. Attachments show up as inline chips in the composer, get expanded into the prompt at send time, and let you drop folders / images / GitHub URLs straight into the conversation.

This is part of the Agent Chat (Beta) surface. With Beta off, attachments are not available.

Attachment Types

Type	Trigger	What it injects
File	`@`, `+` picker, drag-drop	Contents of the file, wrapped with a path header
Folder	`@`, `+` picker, drag-drop	A tree summary + each file's contents, capped at the configured ceiling
GitHub issue	`+` picker (paste issue URL or pick from list)	Title, body, comments, labels, state
GitHub PR	`+` picker (paste PR URL or pick from list)	Title, body, diff, comments, review state
Image	`+` picker, drag-drop, clipboard paste	Sent as a multi-modal message (provider-dependent)

Click the + icon in the composer footer (or press the + icon next to your model picker) to open the unified attachment popup. The popup is organised into modes:

Files — fuzzy-search your workspace tree. Tab to expand a folder for full-folder attachment.
Folders — explicit folder picker for when fuzzy-search is awkward.
Issues — list of open GitHub issues for the current repo, or paste a https://github.com/... URL.
PRs — same as issues but for pull requests.
Images — pick from disk, or use clipboard paste.

The popup dismisses on outside click (Esc also works).

`@` Mentions

Typing @ in the composer opens a fuzzy file picker inline. Select a file and it becomes an attachment chip the same way the + picker does. Faster for "I know the filename" cases.

Drag and Drop

Files / folders from your file manager — drop anywhere on the composer.
Images from your browser or screenshot tool — drop anywhere on the composer.

The drop target highlights when you hover with a valid item.

Clipboard Paste

You can paste images directly into the composer with Ctrl+V (or the platform equivalent). This works for:

Screenshots from grim, the GNOME Screenshot tool, Snipping Tool, etc.
Images copied from a browser (right-click → Copy Image).
Images on your clipboard from any other source.

The paste also works on the new-workspace prompt input, so you can drop a screenshot straight into the "What do you want to do?" textarea when creating a workspace.

Inline Chips

Attached items show up as compact chips in the composer footer, above the textarea. Each chip shows:

Type icon — file, folder, issue, PR, or image thumbnail.
Label — filename, folder name, #42 for issues, PR #137 for PRs.
Tooltip on hover — full path or URL.
Expand button — preview the content that will be injected.
Remove button — drop the attachment.

Send-Time Injection

When you hit Send, attachments are expanded into the underlying API request:

Files and folders — wrapped with path headers; folders walk the tree skipping .gitignore-d entries, hidden files (toggleable), and binary files.
Issues and PRs — fetched fresh from the GitHub CLI at send time; you get the current state, not a snapshot from when you attached.
Images — sent as multi-modal content blocks where the provider supports it (Claude vision, GPT-4o vision). Providers without vision support reject image attachments with a clear error.

Caps and Safety

Per-attachment size cap to prevent a 2-megabyte minified bundle from blowing through your context window.
GIF guard — large animated GIFs are rejected with a friendly warning instead of silently failing the request.
Total context cap — if your attachments would exceed the model's context window, you get a warning before send.

Agent Chat — the chat pane these attachments live in
Workspaces — the new-workspace prompt input also accepts image paste
Pull Requests — the source of the PR data attachments

Attachments

On this page