本文へ移動
cccskills
無料GitHub で公開

design

Explore competing approaches through independent agents before selecting an implementation.

インストール方法を見る

含まれるファイル(2)

  • SKILL.md2.6 KB
  • agents/openai.yaml43 B

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Explore $ARGUMENTS and produce approaches with distinct core mechanisms. Use the smallest useful set of independent explorers; add one only to cover a material alternative or unresolved gap. If delegation is unavailable, disclose that limit and explore sequentially without claiming independence.

Establish the raw problem, actual constraints, and obvious solutions. Reframe where it exposes a different mechanism; do not force reframing for a bounded problem. Give explorers the same relevant repository context, their chosen method, and a distinct question or provocation. They must not see each other's proposals. Reframings are optional starting points, not constraints. Send revisions back to their authors.

Choose methods that expose different mechanisms; this is a menu, not required coverage:

MethodApproach and check
Build upwardStart with the smallest useful unit; check emergent needs and necessary upfront investment.
Work backwardDerive prerequisites from the ideal outcome; check the path is buildable.
Constraint firstSolve or dissolve the hardest constraints; preserve the user outcome.
Transfer an analogyBorrow a mechanism from another domain; identify where the analogy breaks.
EliminateRemove parts while preserving value; account for interactions and useful polish.
AdversarialDesign around consequential failures; state the cost of robustness.
ComposeCombine proven primitives; account for glue and dependency assumptions.
Evolve over timeShow the initial design, observable change triggers, and migration path; avoid speculative infrastructure.
Stakeholder trade-offsExpose conflicting needs and selectable compromises; prioritize rather than satisfy everyone equally.

Each explorer returns its mechanism, benefits, sacrifices, and conditions for choosing it. Compare mechanisms, merge cosmetic variants, and identify shared assumptions. Reopen exploration only when convergence or an uncovered assumption hides a useful option.

Save docs/designs/<short-name>.md with the problem and useful reframings, a landscape of approaches, their trade-offs and selection conditions, shared assumptions, and consequential unexplored territory. Include diagrams when they explain architectural or spatial choices. Stop after presenting the alternatives; implementation requires the user's selection or authorization.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

audit

無料

Run an explicitly selected code or UX audit, or a full fourteen-dimension sweep.

日本語の概要は準備中です。原文の説明を表示しています。

JHostalek/dotclaude112026年9月10日 更新

Apply personal Git conventions when committing, creating or updating PRs, merging, rebasing, or explicitly shipping work end to end.

日本語の概要は準備中です。原文の説明を表示しています。

JHostalek/dotclaude112026年9月10日 更新

judge

無料

Obtain an independent review of the approach and trade-offs in completed work.

日本語の概要は準備中です。原文の説明を表示しています。

JHostalek/dotclaude112026年9月10日 更新

plan

無料

Produce a self-contained HTML implementation plan for explicit approval before coding.

日本語の概要は準備中です。原文の説明を表示しています。

JHostalek/dotclaude112026年9月10日 更新

sota

無料

Produce an up-to-date, state-of-the-art recommendation using domain expertise and brief web research. Use only when explicitly invoked.

日本語の概要は準備中です。原文の説明を表示しています。

JHostalek/dotclaude112026年9月10日 更新

sparring

無料

Examine a plan, decision, belief, or problem through question rounds that explore assumptions, alternatives, and consequences before acting.

日本語の概要は準備中です。原文の説明を表示しています。

JHostalek/dotclaude112026年9月10日 更新

JHostalek のスキルをすべて見る

このスキルの問題を報告する