-
Notifications
You must be signed in to change notification settings - Fork 0
IHaveAdhd
i-have-adhd is a plugin for Claude Code that restructures the assistant's responses to put the answer first and strip out surrounding noise [i-have-adhd repository]. It is presented as ADHD-friendly output, but its author notes no diagnosis is required and the format suits anyone who wants direct, scannable answers [i-have-adhd README].
A default assistant response often opens with preamble, restates the question, adds caveats and tangents, and closes with a pleasantry. The reader has to hunt through that padding to find the part they can act on. For anyone working quickly, or anyone who loses the thread across long replies, the useful content is buried.
The plugin is installed through the Claude Code plugin marketplace and invoked
with the /i-have-adhd command; it can be turned off again with
claude plugin disable [i-have-adhd README]. Once active it applies a
set of formatting rules that reshape every reply: state the answer before any
explanation, break multi-step work into numbered steps, end with one concrete
next action, drop side discussions, recap the current state each turn, give
time estimates in concrete minutes rather than vague terms, show what has been
accomplished, report errors factually, cap lists at five items, and remove
closing phrases such as "Hope this helps" [i-have-adhd README].
The behavior is defined in a single SKILL.md rule file. Changing how the
plugin behaves means editing that file by hand; there is no configuration
interface [i-have-adhd README].
i-have-adhd sits in Tools at Assess, positioned inner. It is a single-purpose output-styling skill for Claude Code, MIT-licensed and widely starred [i-have-adhd repository]. Popularity is not evidence of production use, so the ring stays at Assess. The inner position reflects that it is directly usable in a Claude Code workflow, low-risk to trial, and easy to disable. Its placement moves only after first-person use confirms the reshaped output holds up across real sessions rather than short examples.