New version 0.5.7

You have 0.5.5. Found when this page was made.

What's new:

Every draft error is listed at once; a stdin draft is kept for line edits
A table with no |---| row is an error instead of plain text
Long flow edges are no longer cut off at the edge of the diagram

Run in a terminal:

npx skills update answer-me-with-html -y

Or tell your agent:

Update answer-me-with-html

Claude Duet review: claim check

A

What it means for Allternit

The review supports the bot design we already use. The person's complaints match gaps that Allternit has closed already or can close with a small change.

B

Item by item

Duet complaint	Allternit today (main)	Action
No usage bucket or breakdown	Settings > Usage shows usage per brain and per bot, with a daily chart	Use it in marketing. Bots that run on a Claude plan still share that plan, the same as Duet.
No quiet hours with severity	Quiet hours exist, but they apply to the full account. Only calls get through. Other alerts wait for one summary.	Build it. Add a severity level to each alert, a list of what gets through, and escalation from push to call.
Cannot open or watch the cloud browser	Bot computer viewport and an agent cursor	Differentiator. Make sure the person can open it at any time.
Repeated approvals	Autonomy levels for each place and permission envelopes for each bot	Add a grant for one task: approve the plan one time, and that covers the sends and uploads in it. Payments still ask every time.
Must connect GitHub on claude.ai	Connections are in the app	Differentiator.
Channels like Slack war rooms	Bot threads, projects and the Al coordinator	Check the war-room flow: open, rename when the goal changes, archive, and drop old reminders.

C

Do not claim

Do not say we triage email better than Duet. I did not test this.
Do not say we have "unlimited" usage. A bot on a Claude plan uses that plan.

# Claude Duet review: claim check

## What it means for Allternit
The review supports the bot design we already use. The person's complaints match gaps that Allternit has closed already or can close with a small change.

## Item by item
| Duet complaint | Allternit today (main) | Action |
|---|---|---|
| No usage bucket or breakdown | Settings > Usage shows usage per brain and per bot, with a daily chart | Use it in marketing. Bots that run on a Claude plan still share that plan, the same as Duet. |
| No quiet hours with severity | Quiet hours exist, but they apply to the full account. Only calls get through. Other alerts wait for one summary. | Build it. Add a severity level to each alert, a list of what gets through, and escalation from push to call. |
| Cannot open or watch the cloud browser | Bot computer viewport and an agent cursor | Differentiator. Make sure the person can open it at any time. |
| Repeated approvals | Autonomy levels for each place and permission envelopes for each bot | Add a grant for one task: approve the plan one time, and that covers the sends and uploads in it. Payments still ask every time. |
| Must connect GitHub on claude.ai | Connections are in the app | Differentiator. |
| Channels like Slack war rooms | Bot threads, projects and the Al coordinator | Check the war-room flow: open, rename when the goal changes, archive, and drop old reminders. |

## Do not claim
- Do not say we triage email better than Duet. I did not test this.
- Do not say we have "unlimited" usage. A bot on a Claude plan uses that plan.