shape-idea
Chat and Claude Code: Step 3 of the product-discovery journey: plays the idea back with options, trade-offs and a recommendation. Use when the journey directs it to step 3, not directly.
- License 许可
- MIT
- 所属仓库
- Eliyce/paqad-ai
- 最近更新
- 2026-11-13
- Stars
- ★ 8
- Forks
- 0
01 前言 Front Matter
- name
- shape-idea
- description
- Chat and Claude Code: Step 3 of the product-discovery journey: plays the idea back with options, trade-offs and a recommendation. Use when the journey directs it to step 3, not directly.
- compatibility
- claude.ai chat, Cowork or Claude Code; reads project docs and code through a connector, uploaded files or the open folder
02 资源文件 3
- 参考文档references/challenge-guide.md2.0 KB
- 参考文档references/expert-lenses.md2.0 KB
- 参考文档references/playback-template.md2.5 KB
03 SKILL.md 正文
shape-idea
How to talk: follow references/voice.md. Short, answer first, plain business words, no narration,
one short sources line. Those rules win.
When this runs
Directed by the product-discovery journey, not started directly: step 3, after the idea and today blocks are in the record. Say the status on entry ("Step 3 of 6: shaping the idea.").
What it does
Shape the idea, do not just repeat it, working only from the idea and today blocks and what the user
said. Raise conflicts kindly (references/challenge-guide.md).
- When the today block carries
comparisonmarks, open the playback with them, before any question: the agree, tickets only, repository only and differ lines, one line each, with the wording from step 2. With no marks show no comparison lines and no empty heading; in a tickets-only run label each fact as from the tickets. Then play the idea back, goal first. Use the shape inreferences/playback-template.md: the goal, what exists today (including what the docs could not confirm), the options with their trade-offs in business terms, a smaller first step where it helps, and a recommendation. Call an area the documents do not cover not documented; never say or imply it is not built. Code lines keep "the code suggests" ("What the documents prove" inreferences/paqad-docs-map.md). Ticket lines keep "the tickets say" and their issue keys (references/ground-from-tracker.md). Evidence stays at its level, with the run's one source line (references/evidence-levels.md). End with the sources line, then the closing question through the choice prompt. - Look through the expert lenses (
references/expert-lenses.md): customer value and experience, business and cost, operations and support, data and privacy, security and access, and rollout and change for existing users. Raise only what matters, in business terms, each with a recommendation. Name a privacy or security risk (such as customer personal data leaving the product) plainly with what you would do about it. - Offer options through the host's choice prompt, each with a few-word trade-off (time, cost, risk, effect on users), the recommended one marked with a short why.
- When the idea is large or uncertain, offer a smaller first release or an experiment with a success condition that would justify going further.
- If the idea already exists as a documented capability, or a code-suggests line in the today block shows it exists, say so with the source (keeping "the code suggests" on code lines), recommend the already-exists outcome and draft nothing more unless asked. A ticket line alone never gives that outcome, even for a ticket marked done: say "the tickets say this was finished" with the key, and recommend checking with the team before building again.
- Raise a disagreement between the code or docs and the tickets once, as a challenge with a recommendation that follows the code and docs, never the tickets; the differ line names the issue key and ends "I'm going with the code and docs."
- Label every suggestion as one; nothing enters the record until accepted. Respect an informed choice; revisit it only for new evidence.
If the checks cannot run
No script; shape by hand with this checklist:
- Playback: goal, today, options and trade-offs, smaller first step where useful, recommendation, sources line, closing question last.
- Options and the way forward on a challenge are offered through the choice prompt (a short list in the reply only when the host has none), the recommended one marked.
- Each of the six lenses was considered; the ones that matter for this idea are raised in business terms with a recommendation, the rest left out.
- A real risk, conflict or duplicate is named with its evidence and a way forward.
- Where the idea is large or uncertain, a smaller first release or experiment with a success condition is offered.
- If the idea already exists, that is said with the source and the already-exists outcome is recommended. A ticket line alone, even a done ticket, never gives that outcome: recommend checking with the team first.
- Ticket lines keep "the tickets say" and their keys. A disagreement between the repository and the tickets is raised once, and the repository wins.
- "What the documents prove" (
references/paqad-docs-map.md) by hand: call an area the today block does not document not documented; never say or imply it is not built. Code lines keep "the code suggests". - Every suggested measure or scope cut is labelled a suggestion and only recorded once accepted.
Output
The content of the playback block: goal-first playback, lens notes, options with trade-offs, the smaller first step and the recommendation. product-discovery writes it once the user accepts the shaping.
04 同仓库更多 Skills
start
Chat and Claude Code: One place to start for product work, routing you to the capability that fits. Use it when you want help with product work and a single place to begin ("help me with this ticket", "I have an idea", "what can you do here?", "where do I start?"). It reads what you are after, names the capability that handles it, and hands you straight to it behind the scenes, so you only have to remember one entry. It is a thin router: it does no drafting, keeps no record, and writes nothing to the tracker itself; the capability it hands you to does that work.
ask
Chat and Claude Code: Answers business questions about the product from the project's own documentation and code, in plain words with sources. Use it when someone asks what the product is, what it does or what areas it has ("what is this product?", "what does this project do?", "what areas does it have?"), when a repository is attached or connected and the user asks what it is or what it does, or when they ask how something works today ("can customers pause a subscription today?", "who uses the export screen?", "which orders go to the finance system?"). It works on any repository at the best evidence level it can reach, reads only what the question needs, leads with the answer, adds one short Sources line and one short source line for the level, shows the project's own term on first use, and says honestly where the documents cannot confirm something. It never writes to the tracker, and it can offer to carry a question into the idea journey when the answer reveals a gap.
front-door
Chat and Claude Code: One place to start for product work, routing you to the capability that fits. Use it when you want help with product work and a single place to begin ("help me with this ticket", "I have an idea", "what can you do here?", "where do I start?"). It reads what you are after, names the capability that handles it, and hands you straight to it behind the scenes, so you only have to remember one entry. It is a thin router: it does no drafting, keeps no record, and writes nothing to the tracker itself; the capability it hands you to does that work.
ask-project
Chat and Claude Code: Answers business questions about the product from the project's own documentation and code, in plain words with sources. Use it when someone asks what the product is, what it does or what areas it has ("what is this product?", "what does this project do?", "what areas does it have?"), when a repository is attached or connected and the user asks what it is or what it does, or when they ask how something works today ("can customers pause a subscription today?", "who uses the export screen?", "which orders go to the finance system?"). It works on any repository at the best evidence level it can reach, reads only what the question needs, leads with the answer, adds one short Sources line and one short source line for the level, shows the project's own term on first use, and says honestly where the documents cannot confirm something. It never writes to the tracker, and it can offer to carry a question into the idea journey when the answer reveals a gap.
write-tickets
Chat and Claude Code: Step 5 of the discover journey: drafts epics and user-story tickets with Given/When/Then acceptance criteria into the ticket set and a preview. Also redrafts the ticket set and its preview when the person changes the draft during the journey, for example moving a group of tickets to a different epic or a new one, and asks for approval again. Use when the journey directs it to step 5 or back to it for a redraft, not directly.
shape-idea
Chat and Claude Code: Step 3 of the discover journey: plays the idea back with options, trade-offs and a recommendation. Use when the journey directs it to step 3, not directly.