メニュー
Detect and name hyperfocus sessions that have gone deep on a rabbit hole. Surface the original goal and offer a structured way back.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Manually route the current request through the installed neurodivergent-support skill pack. Use only when the user explicitly invokes `/nd` in Claude Code, `$nd` in Codex, selects ND Router from the skills UI, asks to activate ND support for the current task, or requests ND status/off controls. Select the smallest relevant subset of sibling `nd-*` skills and continue useful work instead of merely describing them.
When presenting options or a backlog, show only the top 3. Hidden items are noted but not shown until one choice is completed.
Before starting a new task thread, surface any tasks that are more than halfway complete. Ask whether to close one first. One-tap dismiss. Never blocks.
Detect complex or multi-step requests and suggest structured planning before acting. Prevents diving into implementation before the problem is understood.
Stabilize task execution when the user explicitly reports frustration, shame, panic, anger, rejection, or overwhelm that is affecting decisions. Preserve work, separate observable facts from interpretations, lower immediate decision load, and choose one reversible next action without turning the exchange into therapy.
Redesign the task environment when a known intention repeatedly fails at the point of action. Use when reminders arrive too early or late, tools or state are hard to find, choices or steps create friction, a system depends on memory, or repeated failure shows that effort alone is not enough.
| name | nd-hyperfocus-guard |
| description | Detect and name hyperfocus sessions that have gone deep on a rabbit hole. Surface the original goal and offer a structured way back. |
Hyperfocus: the ADHD ability to become deeply absorbed in a task to the exclusion of everything else, including the original goal. Hyperfocus is a double-edged strength — it enables deep work, but it can direct that depth at the interesting thing rather than the necessary thing.
A session can spend 4 hours solving an elegant architectural problem that wasn't the actual task. The work is real and good — but it wasn't what needed to happen. The hyperfocus guard names this when it happens.
Apply when:
Name the hyperfocus session (once, gently):
"💡 Deep work check: we've been in [topic area] for [N] turns. The original goal was: [goal]. [Topic area] is interesting work — want to continue here, or would it help to surface back to [goal]?"
Surface the original goal: State it exactly as it was defined, not paraphrased.
Offer a clear branch:
On 'bookmark and surface': Create an explicit save point:
"Bookmarked: [current deep work topic] — [one-sentence summary of what was learned/done]. Now returning to [goal]."
Do not interrupt until turn 10+. Hyperfocus early in a task is often productive depth, not a rabbit hole. The pattern becomes relevant when it persists.
Never moralize. Hyperfocus is not a failure. The guard is purely navigational.
Hyperfocus in ADHD is a real phenomenon: when the dopamine system engages with an interesting or novel task, sustained attention becomes available in a way that's often unavailable for routine work. This is why ADHD is described as a disorder of inconsistent attention, not insufficient attention.
The challenge is that hyperfocus is task-selected by the brain's interest/novelty system, not by the person's conscious goals. The guard creates a brief meta-cognitive check-in: the human decides whether the deep work is serving their goals or displacing them. Most of the time, with the check-in, the human will know the answer immediately.
