Destructive Action Styling
Reserve danger styling for the actions that genuinely destroy something.
What it adds
A single consistent treatment for actions that delete, revoke, or irreversibly change data, applied nowhere else.
What your agent is told to do
5
What your agent is told to do
5-
1
Audit every button, menu item, and icon control in the app and mark which ones destroy, revoke, or irreversibly change something. Only that set earns the destructive treatment.
-
2
Distinguish the two kinds of removal in wording and in weight. Archiving or deactivating something recoverable is not the same act as permanently deleting it, and they must not share a visual.
-
3
Keep destructive actions out of the primary position. They belong in a secondary placement or a menu until the user has chosen them, at which point the confirmation may present the destructive action as its primary control.
-
4
Do not signal danger with colour alone. Pair the colour with wording that names the object and the consequence, and with an icon or a placement difference, so the meaning survives greyscale, colour blindness, and a forced-colours theme.
-
5
Do not let the destructive palette spread into error messages, validation text, and required-field markers — if everything urgent is red, nothing is. Which actions qualify and what the treatment means belongs here; the rest, hover, active, focus, and disabled variants of those same controls belong to the Hover Active Focus Checklist.
Edge cases it handles
7
Edge cases it handles
7- A soft delete that can be undone from a trash view must not look identical to a permanent purge, or users learn to fear both equally and hesitate over the harmless one.
- Colour alone fails for anyone with a colour vision deficiency, and it fails entirely in a high contrast theme that overrides the palette with system colours.
- A destructive action rendered as the loudest control on the screen gets clicked by accident. It should stay quiet until the user has committed to it inside a confirmation.
- The destructive colour has to meet contrast requirements on hover, while active, while focused, and while disabled, not only at rest — muted danger buttons routinely fall below the threshold once they are dimmed.
- Bulk destructive actions need the affected count in the label and in the confirmation, because deleting one row and deleting four hundred look identical at the moment of the click.
- In a dark theme the usual danger red vibrates against a near-black surface. Pick a value tuned for the dark palette rather than reusing the light one.
- An action that is destructive for other people — revoking a colleague's access, removing a shared record — deserves the same treatment as destroying your own data.
Definition of done
8
Definition of done
8- Only actions that delete, revoke, or irreversibly change data carry the destructive treatment.
- Recoverable removal and permanent removal are distinguishable both visually and in the wording.
- Destructive intent is carried by wording and placement as well as colour, and survives in greyscale.
- No destructive control is the visually dominant action on a screen outside a confirmation.
- The destructive palette meets contrast requirements at rest, on hover, on focus, and when disabled, in both themes.
- Error and validation styling is visually distinct from destructive action styling.
- The feature matches the existing design system.
- No existing functionality is broken.
Related features
Feedback Widget
Feedback Widget
Let anyone report a bug or send a comment from the page they are looking at.
What it does
An always-available launcher that opens a short feedback form and submits it with page context attached.
How it works
- 1 Put a single persistent launcher in one corner of the app and open a compact form in place. Do not route the user to a separate feedback page and lose the context they were about to describe.
- 2 Attach the current URL, the app version, the viewport size, the signed-in account, and a short trail of recent client errors automatically, so the reporter does not have to describe where they were.
- 3 Show the captured context to the user before they send it, in a collapsible section, and let them remove it. Silent collection of screen state is a trust problem.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/feedback-widget
Audio Chapter Markers
Audio Chapter Markers
Break a long recording into named chapters listeners can see and jump between.
What it does
Named, timestamped chapters on long audio or video, shown in a list and on the scrubber, each addressable by its own link.
How it works
- 1 Store chapters as an ordered list of start times and titles attached to the existing media record, and render them both as a list beside the player and as markers on the scrubber.
- 2 Let an author add, rename, reorder by moving start times, and delete chapters in the player itself, with the current playback position offered as the default start time for a new chapter.
- 3 If chapters are generated automatically, treat the generated set as a first draft. The author's edits must win and must survive the next generation run.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/audio-chapter-markers
Interactive Transcript
Interactive Transcript
Put the transcript next to the player so clicking any line jumps playback to that moment.
What it does
A time-aligned transcript panel beside the audio or video player, with click-to-seek, a moving highlight, and in-transcript search.
How it works
- 1 Render the transcript from the timestamps the app already stores. If Voice Transcription exists, extend its output and its data model rather than producing a second transcript with its own timings.
- 2 Make every line a real seek target: clicking or pressing enter on a line moves playback to that time and keeps focus where the user left it.
- 3 Highlight the line currently being spoken and keep it in view as playback advances, using the player's own time updates rather than a separate timer.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/interactive-transcript
How it works
-
1
Copy the link
Grab the Markdown instruction URL for this feature.
-
2
Give it to your AI
Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.
-
3
It inspects, then implements
Your agent reads your existing app first, then adds the feature to fit it.
Works with your stack
These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.
Need it tighter than that? Customize the feature and tell it exactly what you're running.