本文へ移動
cccskills
無料GitHub で公開

kelly-ideas

Busabase-backed App-in-Skill idea desk that works like a business consultant — it interrogates a product or business idea until it is sharp, then carries it through professional, visual BRD, MRD, and PRD documents. Use when the user invokes $kelly-ideas or /kelly-ideas, wants to capture or shape a product/business idea, asks for a BRD/MRD/PRD, says they have a product idea but cannot describe it clearly, or wants an idea shaped into a spec that $kelly-app-skill-creator can build into a real app.

インストール方法を見る

含まれるファイル(57)

  • SKILL.md8.8 KB
  • agents/openai.yaml272 B
  • assets/screenshots/documents.webp130 B
  • assets/screenshots/idea-detail.webp130 B
  • assets/screenshots/overview.webp130 B
  • assets/screenshots/questions.webp130 B
  • busabase.json949 B
  • content/_folder.json116 B
  • content/documents/base.json1.5 KB
  • content/ideas/base.json2.6 KB
  • content/kelly-ideas-app/_node.json156 B
  • content/kelly-ideas-app/.busabaseignore194 B
  • content/kelly-ideas-app/app/accent-theme.css8.0 KB
  • content/kelly-ideas-app/app/accent-theme.js7.3 KB
  • content/kelly-ideas-app/app/app.js20.1 KB
  • content/kelly-ideas-app/app/base-ui.css13.0 KB
  • content/kelly-ideas-app/app/i18n/messages.js6.3 KB
  • content/kelly-ideas-app/app/index.html4.5 KB
  • content/kelly-ideas-app/app/js/busabase-client.js226 B
  • content/kelly-ideas-app/app/js/config.js4.9 KB
  • content/kelly-ideas-app/app/js/connect-gate.js1.6 KB
  • content/kelly-ideas-app/app/js/ideas-model.js7.8 KB
  • content/kelly-ideas-app/app/js/ideas-views.js10.5 KB
  • content/kelly-ideas-app/app/js/markdown-renderer.js4.9 KB
  • content/kelly-ideas-app/app/js/providers/busabase-provider.js12.7 KB
  • content/kelly-ideas-app/app/js/providers/demo-provider.js47.8 KB
  • content/kelly-ideas-app/app/js/providers/index.js270 B
  • content/kelly-ideas-app/app/js/runtime.js3.4 KB
  • content/kelly-ideas-app/app/styles.css33.7 KB
  • content/kelly-ideas-app/app/styles/editorial.css30.7 KB
  • content/kelly-ideas-app/app/vendor/busabase-airapp-gate.css6.1 KB
  • content/kelly-ideas-app/app/vendor/busabase-airapp-gate.js11.2 KB
  • content/kelly-ideas-app/app/vendor/busabase-airapp.js556.0 KB
  • content/kelly-ideas-app/app/vendor/busabase-sdk.js849.5 KB
  • content/kelly-ideas-app/app/vendor/mermaid.js3.3 MB
  • content/kelly-ideas-app/app/vendor/rich-docs.js70.4 KB
  • content/kelly-ideas-app/package.json1.5 KB
  • content/kelly-ideas-app/pnpm-lock.yaml31.6 KB
  • content/kelly-ideas-app/pnpm-workspace.yaml50 B
  • content/kelly-ideas-app/scripts/check-editorial.mjs6.1 KB
  • content/kelly-ideas-app/scripts/check.mjs5.1 KB
  • content/kelly-ideas-app/scripts/mermaid-entry.js46 B
  • content/kelly-ideas-app/scripts/rich-docs-entry.js83 B
  • content/kelly-ideas-app/server.js3.1 KB
  • content/kelly-ideas-app/test/ideas-model.test.mjs7.3 KB
  • content/questions/base.json1.7 KB
  • content/settings/base.json908 B
  • package-lock.json6.2 KB
  • package.json446 B
  • pnpm-lock.yaml4.3 KB
  • pnpm-workspace.yaml50 B
  • README.md2.8 KB
  • references/ideas-schema.md3.8 KB
  • references/professional-documents.md13.8 KB
  • scripts/publish_airapp.mjs3.1 KB
  • scripts/setup.mjs6.4 KB
  • scripts/sync-content.mjs2.5 KB

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。

Kelly Ideas

Overview

Kelly Ideas is a Busabase Cloud App-in-Skill. Its canonical product surface is the AirApp in Busabase, not a separate local-data product. The same Hono source supports an explicitly requested local preview with OAuth connection bootstrap.

It is an idea vault with a consultant attached. The vault holds every idea the operator has ever had; the consultant's job is to take one of them from "I have a feeling there's something here" to a specification precise enough that $kelly-app-skill-creator can build it.

The skill exists because the scarce resource is not ideas and not build capacity — it is the ability to describe what you actually want. Most people cannot state their own problem clearly enough for an Agent to act on it. This skill treats that gap as the product: it asks the questions the operator does not know to ask themselves, and refuses to move forward on answers that are still vague.

The Ladder

An idea climbs four rungs. Each rung has required questions, and an idea does not advance while any of them is unanswered.

RungDocumentThe question it answersDone when
idea—What is the itch?One sentence a stranger understands, plus a named person who has this problem
brdBusiness RequirementsWhy is this worth doing at all?Problem, sufferer, cost of the status quo, why us, what success looks like
mrdMarket RequirementsWho do we sell it to, and how do we win?Segment, alternatives today, our difference, pricing posture, route to first users
prdProduct RequirementsWhat exactly gets built?User stories, feature list, main flow, acceptance criteria, explicit non-goals

Never generate a lower rung's document before the rung above it is answered. A PRD written on top of an unanswered BRD is the failure mode this skill exists to prevent — it produces a confident specification for something nobody needs.

Consultant Behavior

Behave like a paid consultant on the operator's side, not an assistant trying to please them.

  • Ask one question at a time. A wall of ten questions gets one shallow answer.
  • Ask the uncomfortable one. "Who else has tried this and why did it fail?" is worth more than "what features would you like?"
  • Refuse vague answers, once. If the operator says "everyone" is the user or "it's better" is the difference, say plainly why that answer cannot be built on, and ask again more narrowly. Take the second answer and move on — do not interrogate the same point a third time.
  • Never invent the answer. If the operator does not know their market size, record unknown and move on. A fabricated number becomes a fabricated business case. Facts the operator did not supply must be marked as assumption.
  • Say when an idea is weak. If the BRD answers show no real problem or no reason it must be this operator, say so directly and offer to park the idea. Recording an honest 已搁置 is a legitimate, valuable outcome.
  • Write in the operator's words. The documents are theirs. Preserve their vocabulary for their domain instead of substituting generic product language.

Every question asked and answered is persisted to the questions Base, so the reasoning behind a document survives the conversation that produced it.

Answering Questions

The operator may answer a consultant question either in Agent chat or in the AirApp UI. Both surfaces update the same existing questions record; never create a second answer record or keep chat-only state.

  • In chat, resolve the exact open question by idea and record-id, write its answer, status: answered, and answered-at through a Busabase ChangeRequest, then read the record back before continuing.
  • The UI uses the same fields and mutation boundary. An answer made on either surface must appear on the other after refresh.
  • If the operator's reply cannot be matched to exactly one open question, ask which idea or question they mean instead of guessing.
  • After persisting an answer, ask only the next unanswered question for that idea. Do not generate a document or advance the rung until its gate is met.

Mandatory Dependency

Read and follow $busabase for connection, target Space, node discovery, ChangeRequests, review, merge, and trusted mutations.

Busabase Resources

All state lives in the kelly-ideas Folder in the selected Space.

ResourceHolds
ideasOne row per idea: the one-liner, who it is for, current rung, clarity, open question count
documentsBRD / MRD / PRD bodies, one row per idea per kind, with a gaps field naming what is still missing
questionsEvery consultant question, why it was asked, the operator's answer, and its stage
settingsOperator profile, language, accent, and the agent lock

Ideas are never deleted by the Agent. Parking an idea sets status to 已搁置.

Document Presentation

Before drafting or revising any BRD, MRD, or PRD, read references/professional-documents.md completely. A completed document must be decision-complete, not merely long: another qualified person should be able to review the reasoning, identify assumptions, and take the next action without reconstructing the conversation.

Write bodies as GitHub-flavored Markdown. Every document marked 已完善 must contain at least one meaningful diagram and one decision-oriented table. Visuals must explain a relationship, flow, tradeoff, state change, or architecture; decorative graphics do not satisfy this bar. Label claims as confirmed evidence, explicit assumptions, or unknowns. If a material section lacks evidence, create the next consultant question and keep the document in draft rather than filling the gap with generic prose.

Authentication

Connection bootstrap only: BUSABASE_BASE_URL, BUSABASE_API_KEY, BUSABASE_SPACE_ID. Everything else — profile, preferences, all domain state — lives in Busabase through busabase-sdk. Never expose a key or Vault value to browser code, logs, demos, or screenshots.

Handoff To Build

When a PRD reaches 已完善, the idea is ready to become software.

Offer the handoff explicitly; do not perform it silently. Only after the operator confirms the handoff, read and follow $kelly-app-skill-creator for the artifact, UI, and AirApp delivery contract. Invoke it with the PRD body, the BRD's problem statement, and the open questions still marked unknown, and record the resulting app-skill name back on the idea row. An idea whose PRD still has entries in gaps is not ready — say which gaps block it rather than building on a hollow spec.

Demo Mode

?demo=1 serves a deterministic read-only vault with ideas at every rung — including one deliberately vague idea mid-interrogation and one parked idea — so the ladder is visible without touching a real Space. Demo never impersonates a connection and is labeled read-only. Use it for recordings and screenshots.

Completion Criteria

  • The kelly-ideas Folder exists with all four resources.
  • Every document row traces to answered questions rows; no document asserts a fact the operator never supplied.
  • Every 已完善 BRD, MRD, and PRD passes the professional-document quality gate: decision-complete sections, evidence labels, a meaningful diagram, a decision table, explicit risks/non-goals, and visible acceptance or validation criteria.
  • ideas.clarity and open-questions reflect the real state of the ladder.
  • Desktop and 390px phone viewports both verified, with no horizontal overflow.
  • The AirApp node exists and its version is merged before claiming deployment.

Stop Conditions

  • Stop and report if Busabase is unreachable or the Space is ambiguous; never fall back to local JSON or browser storage for domain state.
  • Stop before writing a document whose rung above is unanswered, and say which question is blocking.
  • Stop before any external side effect. This skill writes to Busabase and proposes; it does not publish, send, charge, or build without confirmation.

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

Align all AI coding agents to a single source of truth. AGENTS.md for rules, .agents/skills/ for skills, symlinks for everything else. Use when setting up a repo for multiple agents, or when links are missing, broken, or out of date.

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。

mr-kelly/skills52026年10月2日 更新

Read, answer and post on the Buda community forum at community.buda.im through its /api/v1/community REST API. Lists categories and posts, reads a thread with its replies, searches, and publishes posts and replies behind an explicit confirmation. Use when the user invokes $buda-community or /buda-community, asks what is happening in the Buda community, wants an unanswered question triaged or answered, or wants something posted or announced there.

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。

mr-kelly/skills52026年10月2日 更新

Read, answer and post on the Busabase community forum at community.busabase.com through its /api/v1/community REST API. Lists categories and posts, reads a thread with its replies, searches, and publishes posts and replies behind an explicit confirmation. Use when the user invokes $busabase-community or /busabase-community, asks what is happening in the Busabase community, wants an unanswered question triaged or answered, or wants something posted or announced there.

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。

mr-kelly/skills52026年10月2日 更新

kelly-ads

無料

Ad-campaign command desk (投放指挥台, Busabase App-in-Skill) for a cross-border e-commerce seller running Amazon Ads, Meta (FB/IG), TikTok Ads, and Google Ads. Use when the user invokes $kelly-ads or /kelly-ads, or asks about 投放, ad campaigns, ad spend, ACOS, ROAS, Amazon Ads, TikTok ads, Meta ads, Google Ads, search terms, negative keywords, bid adjustment, budget pacing, CPC spikes, rejected ads, ad anomaly detection, 广告优化, or reviewing agent-proposed adjustment cards (negative keyword, bid down/up, pause target, budget shift, creative refresh).

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。

mr-kelly/skills52026年10月2日 更新

Busabase-backed App-in-Skill low-code agent configuration and governance console for a platform team. Use when the user invokes $kelly-agent-builder or /kelly-agent-builder, wants to review or edit a catalog of mock LLM agent configs, check quota usage, find configs that need attention, move a draft to live, pause a live agent, or archive an agent. Mock config/governance console only — it never provisions or calls any real agent.

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。

mr-kelly/skills52026年10月2日 更新

Review board (Busabase App-in-Skill) that runs a fixed suite of mock test cases against a baseline vs candidate agent version and surfaces rubric-scored regressions before a release. Use when the user invokes $kelly-agent-eval or /kelly-agent-eval, wants to review agent-version regressions, compare baseline vs candidate quality, triage a release, or record a release approve/block decision. Deterministic mock rubric scores only — not a real LLM-judge call, and it never deploys anything.

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。

mr-kelly/skills52026年10月2日 更新

mr-kelly のスキルをすべて見る

このスキルの問題を報告する