decide-one-at-a-time

Presents queued decisions one at a time with mode-graded explanation depth and discussion gates.

Updated Apr 11, 2026
One-click install
npx skills add https://github.com/gozonerd/shadow-ai-assessment --skill decide-one-at-a-time-gozonerd
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: decide-one-at-a-time
Source: https://github.com/gozonerd/shadow-ai-assessment/tree/main/.claude/skills/decide-one-at-a-time
Command: npx skills add https://github.com/gozonerd/shadow-ai-assessment --skill decide-one-at-a-time-gozonerd

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Large batches of pending decisions presented as a single stacked list overwhelm users and stall progress. This Skill serializes a decision queue into one-at-a-time prompts with a discussion gate between each, so users can rapid-fire decide without cognitive overload. ## Core Features & Use Cases - Four explanation modes: solo (options only), explain (plain-English per option), rec (explanation plus recommendation), and defend (per-option for/against rationale with a defended recommendation). - Structured decision framing: each decision gets a human-readable UPPERCASE-KEBAB-CASE code and a count-out-of-total header (e.g., D3 of 8) for progress tracking. - User pacing overrides: users can bundle remaining decisions, switch modes mid-queue, skip or defer items, and receive a closing summary of resolved and pending decisions. - Use Case: After a project milestone close-out surfaces 8 follow-up decisions, invoke the skill to walk through each decision individually with recommendations instead of facing one dense wall of options. ## Quick Start Ask the AI to walk you through your pending decisions one at a time using /decide-one-at-a-time in rec mode.

Frequently Asked Questions about decide-one-at-a-time

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I present multiple decisions to a user one at a time?

Queue the decisions with codes and enumerated options, confirm scope and mode in a single Step 0 message, then send each decision as its own message and stop until the user answers. The skill enforces a discussion gate between decisions.

What are the solo, explain, rec, and defend modes?

They are four explanation-depth levels. Solo shows options only, explain adds plain-English meaning per option, rec adds a recommendation with reasoning, and defend adds per-option for/against analysis plus a defense of the recommendation.

When should I not use one-at-a-time decision presentation?

Avoid it when only 1-2 decisions are pending, when decisions are tightly coupled so one determines the next, or when the user explicitly wants the full menu at once. Open-ended discussions without enumerable options also do not fit.

Can the user override the one-at-a-time pacing?

Yes. The user can say 'give me all of them at once' or 'cluster the rest' to receive remaining decisions stacked, switch modes mid-queue, skip or defer decisions, or stop early and get a closing summary.

How are decisions labeled and tracked during the queue?

Each decision gets a stable UPPERCASE-KEBAB-CASE code that evokes its substance, plus a count-out-of-total header like 'D3 of 8'. The session ends with a summary table of resolved and deferred decisions.