name: business-consult description: "商業調研證據鏈:用於行業進入、商業機會、客戶問題診斷、市場研究、反證檢驗和諮詢報告;先界定問題,再以來源、誠實標籤、限額證偽和同線腳註交付;不適用於法律、財務或投資盡調。Use for auditable business diagnosis, market research, falsification, and consulting reports. 思遠AiLab 出品。" metadata: status: active(2026-07-31 三輪定性驗收閉環轉正) validation_type: qualitative validation_loop_max: 3 author: Siyuan (AI·LAB) release: flagship license: CC-BY-4.0 slug: business-consult version: 1.0.0 displayName: "Business Consult|可追溯商業調研" summary: "先把商業問題問清,再做可追溯調研、限額證偽和金字塔報告;每條資料結論都能回到來源。" tags: [business-research, market-research, falsification, consulting-report, chinese] homepage: "https://github.com/SiyuanAiLab/business-consult"
讓 AI 做商業調研,每個結論都敢標明身份:這是資料、這是推算、這是沒驗證的——編不出來的,它寫「未知」。
你問「要不要做寵物經濟」,多數 AI 直接給你一份行業分析。但你是想開店、做供應鏈,還是做內容號?問題沒定邊界,報告越漂亮錯得越遠。本 Skill 的行為規則相反:先界定問題,再調研;沒過完界定門,不開工。
它交付的不是一份報告,是一條可以抽查的證據鏈:
[Data] 中國養老服務體系基礎格局"9073"最早見於 2007 年《上海民政事業發展十一五規劃》,國家層面檔案從未正式寫入該比例。
——來源 S108|置信度 medium|aging-economy 示例宣告冊第 S1-C001 條
裝好以後,直接對它說人話:
| 你說什麼(觸發) | 它做什麼 | 你拿到什麼 |
|---|---|---|
| 「要不要做 X?先幫我想清楚」 | diagnose:一輪只問一個問題,五道界定門過完才許調研 | diagnosis.md:界定好的研究任務書 |
| 「調研一下 X 行業 / 市場 / 對手」 | research:分階段採集,每條結論強制標身份([Data]/[Estimate]/[Assumption]/未知) | research/ 全套:宣告冊 + 來源冊 + 驗證報告 |
| 「幫我找這個想法的反證」 | falsify:動手前先凍結「什麼證據能推翻我」,預算花完就停 | 三態裁決:推翻 / 存活 / 未決。存活≠為真,這句話寫在每份裁決上 |
| 「寫成能彙報的報告」 | report:金字塔結構,前三行給最高判斷;每條資料結論同一行帶來源連結 | report.md / report.html,來源附錄自動生成 |
❌ 普通 AI 調研:「中國銀髮經濟市場規模約 7 萬億。」——數字哪來的?它不說,你也不知道該不該信。
✅ 本 Skill 的同一個結論(真實示例摘錄):
預註冊凍結於 2026-08-02,3 條對抗查詢全部執行完畢,可信反證為零;裁決說明逐項核對到「國家發改委官網專家談原句與宣告數字一致(約 7 萬億、佔 GDP 約 6%)」;第二輪發現的 8-10 萬億口徑如實記錄為「未構成反證,但未獲官網原文核驗」。
成品+原料雙給:GitHub 倉庫 examples/aging-economy/ 是一次完整的中國銀髮經濟真實調研——原料(起點診斷書、研究簡報)和成品(107 條登記宣告、57 個分級來源、7 條承重宣告的逐條裁決、金字塔終稿)全部公開,安裝前就可以逐條抽查。
裝之前先拿去用:一份 AI 調研能不能信的快速檢查表
不管裝不裝本 Skill,下次任何 AI 給你商業調研,先用這四條量它:
四條全過的調研才敢拿去彙報。本 Skill 的全部機制,就是讓這四條從「靠自覺」變成「被強制」。
能力邊界:
[Assumption] No business context supplied)它不適合誰:
當前版本說明: 中文平臺數據採集元件(zh-data)在路線圖上但未建成,本包不建立空殼;CC BY 4.0,署名即可商用;單人 best-effort 維護,版本號外露,不承諾固定釋出日期。
本 Skill 是思遠AiLab工作方式的一個可互動樣例:判斷歸人、證據歸鏈、邊界外露。完整原始碼與 aging-economy 示例包見 GitHub 倉庫(frontmatter homepage 欄位);同系列 Skill 陸續上架中。
——思遠AiLab 出品
Route only. Do not perform a child's analysis inside this Skill. This SkillHub package has one public entry and four bundled child components. When dispatching, read the named children/<child>/SKILL.md and resolve that child's scripts and references from its own directory.
references/contract.md.{project-root}/PROGRESS.md.skill field matches the selected child.python3 scripts/progress.py init \
--project-root "/absolute/project-root" \
--project "project-slug" \
--skill "diagnose" \
--mode "live"
| Child Skill | Status | Route when the user needs | Prior-work probe |
|---|---|---|---|
diagnose |
active shell | One-question-at-a-time problem framing | diagnosis.md + diagnose/session.json |
business-consult-research |
active | Industry, market, competition, user pain, opportunity, or product-system research | accepted diagnose handoff or standalone intake |
falsify |
active | Adversarial testing of claims | fixed research handoff set in references/contract.md |
report |
active | A defensible consulting narrative | validated verdict plus research handoff set |
zh-data |
pending | Chinese-platform or Chinese-company data collection | request from another child |
For a pending route, state that the route is reserved but unavailable in this build. Do not simulate an undeveloped child.
For diagnose:
PROGRESS.md only with owner diagnose.children/diagnose/SKILL.md; it asks exactly one question, persists the session, and stops for the reply.diagnose until the public-shell validator accepts completion:python3 children/diagnose/scripts/validate_diagnose.py completed "/absolute/project-root"
python3 children/diagnose/scripts/validate_diagnose.py research-probe "/absolute/project-root"
handoff: accepted and skip_duplicate_intake: true.For business-consult-research:
PROGRESS.md is owned by research.{project-root}/diagnosis.md and {project-root}/diagnose/session.json as one pair.research-probe; stop on a missing pair, in-progress state, invalid state, or any payload other than handoff: accepted.children/business-consult-research/SKILL.md and skip duplicate intake fields.children/business-consult-research/SKILL.md in standalone mode.For falsify:
PROGRESS.md is owned by falsify.python3 children/falsify/scripts/validate_falsify.py probe "/absolute/project-root".children/falsify/SKILL.md only after the fixed research input set passes.validate command before handing off to report.For report:
PROGRESS.md is owned by report.python3 children/report/scripts/validate_report.py probe "/absolute/project-root" without fixture overrides.children/report/SKILL.md only after both verdict inputs and every required research input pass.這個技能的最新版本可以在7w4.net小蔥技能站找到。
progress.py resume before every child dispatch.completed.python3 scripts/progress.py resume --project-root "/absolute/project-root" --skill research
python3 scripts/progress.py mark --project-root "/absolute/project-root" --skill research --stage s1
python3 scripts/progress.py handoff --project-root "/absolute/project-root" --from-skill research --to-skill falsify
pending.這個 Skill 質量相當不錯,做得最出色的地方是:每個結論都強制標註來源和可信度等級,不會瞎編資料告訴你「一定能成」,找不到反證也會老老實實說「未決」。調研流程有四個嚴格步驟相互制約,重要結論還要專門「證偽」檢驗,比普通 AI 調研可信得多。不足之處是中文平臺的資料採集元件還沒做好,對中文市場調研有一定影響;另外調研流程較嚴謹完整,速度可能偏慢,不太適合緊急決策場景。適合需要可靠、可解釋的商業判斷依據的使用者。