Skip to main content

shaping-framing

Create an evidence-backed frame from messy transcripts and notes before capsule/card planning. Use when discovery inputs are conversation-heavy and the problem statement is still diffuse.

インストールへ移動

ソース情報

リポジトリ
hashintel/labs
ソースの最終更新活動
2026年3月23日 12:12
検出された SKILL.md の言語
英語
スター
35
フォーク
14

インストール方法

デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。

ソースファイルを確認

インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。

SKILL.md を表示中

SKILL.md
ソースの指示 · 読み取り専用プレビュー
name
shaping-framing
description
Create an evidence-backed frame from messy transcripts and notes before capsule/card planning. Use when discovery inputs are conversation-heavy and the problem statement is still diffuse.
argument-hint
[topic and transcript paths, or context bundle to frame]
# Shape Framing (Adjunct) Use this adjunct skill to distill conversation artifacts into a clear frame that can feed pragma planning. ## Source And Adaptation Note Adopted from `rjs/shaping-skills` (`framing-doc`) and adapted for this repository. Modifications: - Namespaced as `pragma:shaping-*` to avoid routing ambiguity - Explicit pragma re-entry requirements - Authority boundaries aligned to `docs/shape-pragma-integration.md` ## When To Use - Inputs are transcripts, call notes, or scattered stakeholder messages - The team agrees a problem exists but cannot state it crisply yet - You need an evidence-backed "why now" before capsule/card work ## When Not To Use - You already have a stable capsule and clear next slice - The task is implementation-ready and only needs `/pragma:card` ## Procedure 1. Collect source inputs and confirm ordering when chronology matters. 2. Extract direct evidence and separate it from synthesis. 3. Produce a frame with Problem, Outcome, and boundaries. 4. Identify what changed in understanding and what remains uncertain. 5. Define the pragma re-entry step. ## Required Output ```md ## Frame [evidence-backed problem and outcome] ## Handoff To Pragma - Capsule impacts: [terms/invariants to add or revise] - Candidate cards: [1-3 candidate slice behaviors] - Open uncertainty: [items requiring `/pragma:spike`] - Next command: [`/pragma:capsule` or `/pragma:card`] ``` ## Constraints - Do not prescribe implementation details. - Do not replace capsule authority over vocabulary and invariants. - Do not route directly to `/pragma:slice`. ## Lifecycle - **State**: `discovery` - **Next**: `/pragma:capsule` (default) or `/pragma:card` (if capsule is stable) - **Loop**: `/pragma:consult`
GitHubで見る