ig-reply
Skill19 starsChanged 3 months agoMITView on GitHub
Instagram 貼文留言智慧回覆。抓某篇貼文底下的新留言,由 Claude 在對話裡依貼文內容與留言者語言生成友善回覆,再透過 Graph API 自動送出。觸發時機:使用者說「回覆留言」「處理 IG 留言」「回 IG」「IG 自動回覆」「抓留言來回」「回我 IG 那篇」等。
- Reads credentials
- Installs packages
7 other projects use a version of this.
Adopted by
Most recently taken up by
- aipphilie/claude-skillstheir own version, first seen Nov 1, 2026
- Oumi-Mez/Claude-skillstheir own version, first seen Oct 25, 2026
- lodhaarihant/instagram-agent-arihanttheir own version, first seen Oct 17, 2026
- postzen-dev/instagram-agent-skilltheir own version, first seen Oct 11, 2026
- nessalazne/instagram-agent-skilltheir own version, first seen Sep 28, 2026
- larkinmd7/instagram-agent-skilltheir own version, first seen Sep 27, 2026
Used with
Projects with it also use ig-profile, ig-plan and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
JakeschincariolSkillig-profile
Score an Instagram profile out of 100 against a 12-part rubric and rewrite the parts that lose points - name field, bio, link, highlights, pinned three, grid. Use when the user says "optimize my profile", "fix my bio", "score my Instagram", "why don't people follow me", or pastes their profile and asks how it reads.
Also in 7 projects that use this
JakeschincariolSkillig-plan
Build the week on Instagram - what to post, which format, when, and who to engage with. Use when the user says "plan my week", "what should I post", "content calendar", "I have nothing to post about", or wants a posting schedule and an engagement list.
Also in 7 projects that use this
nicojunkSkillig-comment
Comment response strategy and semi-automated handling. Defines response rules by comment category, maintains the creator's authentic voice, and references Make.com automation blueprints for scalable comment management.
Also in 7 projects that use this
nicojunkSkillig-repurpose
Cross-platform content repurposing from Instagram to TikTok, YouTube Shorts, LinkedIn, Email Newsletter, and Blog. Adapts format, aspect ratio, duration, tone, and CTA per platform while preserving core message.
Also in 7 projects that use this
- +3 more from Jakeschincariol
- +5 more from nicojunk
Other versions7 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 7
- Jakeschincariol/instagram-agent-skill★ 88same name
- postzen-dev/instagram-agent-skill★ 34same name
- nessalazne/instagram-agent-skill★ 3same name
- aipphilie/claude-skills★ 0same name
- larkinmd7/instagram-agent-skill★ 0same name
- lodhaarihant/instagram-agent-arihant★ 0same name
- Oumi-Mez/Claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
Files
ig-reply — Instagram 留言智慧回覆
這個 skill 不會呼叫任何付費 AI API。生成回覆的那一步,由 Claude 在對話裡直接做。 程式只負責兩件事:抓資料 和 送出回覆。
資料夾結構
ig-reply/
├── SKILL.md ← 你正在看的這份說明
├── .env ← 你的憑證(不要分享給任何人,已被 .gitignore 擋住)
├── .env.example ← 範本(可以分享)
├── .gitignore ← 告訴 git 不要上傳 .env、留言資料等
├── package.json ← Node.js 設定
├── oauth.js ← 一次性:走 FB Login 拿 Page Token
├── test.js ← 步驟 A:測試 token 是否有效、列出貼文
├── fetch.js ← 步驟 B:抓某篇貼文的留言
├── sweep.js ← 步驟 B(進階):一次掃「所有貼文」的新留言
├── reply.js ← 步驟 D:把回覆送出去
├── skip.example.json ← 「不回」名單的格式範本(複製成 skip.json 使用)
├── replied.json ← (自動產生)記錄已回過的留言,避免重複
├── pending.json ← (自動產生)fetch.js 寫出,給 Claude 看
├── pending_sweep.json← (自動產生)sweep.js 寫出
└── replies.json ← (你/Claude 填)要送出的回覆,reply.js 會讀第一次使用:拿 Token(只做一次)
本 skill 走 Facebook Login flow(用 Page Access Token 讀 IG 留言),不是 Instagram Login。
- 在 Meta for Developers 建一個 App,拿到 App ID 和 App Secret。
- 複製
.env.example成.env,填入FB_APP_ID與FB_APP_SECRET。 - 執行
npm install,再執行node oauth.js:它會開瀏覽器走 FB 同意流程,自動把FB_PAGE_ACCESS_TOKEN/FB_PAGE_ID/IG_BUSINESS_ACCOUNT_ID/IG_USERNAME寫回.env。- 需要的權限:
pages_show_list、pages_read_engagement、instagram_basic、instagram_manage_comments、business_management - 你的 IG 帳號必須是商業帳號且連結到一個 FB 粉專。
- 需要的權限:
完整流程(每次回覆都重複這 4 步)
步驟 A:測試 token 是否還有效
npm install # 第一次才需要
node test.js成功會印出你最近的幾篇貼文,每篇有編號(media_id)、發佈時間、說明文字前幾個字。 看到錯誤就告訴 Claude(常見錯誤訊息會直接是中文,例如「Page Token 過期,請重跑 node oauth.js」)。
步驟 B:抓某篇貼文的新留言
從步驟 A 的輸出複製想處理的 media_id:
node fetch.js <media_id>程式會抓該篇說明+所有留言(含巢狀回覆)、過濾掉已回過的(依 replied.json)以及你放進
skip.json 的「不回」名單,寫到 pending.json,純表情符號的留言會另外標到 emoji_only 清單。
步驟 B(進階):一次掃「所有貼文」的新留言
不想一篇篇 fetch,可以用 sweep 一次掃出所有貼文裡還沒回的新留言:
node sweep.js # 掃全部歷史貼文
node sweep.js 2024-01-01T00:00:00+0000 # 只掃某日期之後的貼文(較快)結果寫到 pending_sweep.json(結構同 pending,但把各貼文分組)。一樣會套用
replied.json(已回)與 skip.json(不回)兩份名單,不會重複跳出。
skip.json(選用):想把某些留言標成「已讀不回」(垃圾、亂碼、收尾語),把它們的 留言 id 寫進
skip.json(格式見skip.example.json)。fetch.js / sweep.js 掃描時會 自動略過。此檔含你自己的備註,已被.gitignore擋住、不會上傳。
步驟 C:請 Claude 生成回覆
跟 Claude 說「幫我讀 pending.json,生成 replies.json」。Claude 會依留言者語言+貼文主題,
逐則生成友善、簡短、像真人的回覆,寫成 replies.json:
{
"留言ID": "回覆文字",
"留言ID": "回覆文字"
}步驟 D:把回覆送出去
node reply.js讀 replies.json 逐則透過 Graph API 送出,每則之間隨機等 30~90 秒(避免被判定為機器人),
送出成功的留言 ID 記錄到 replied.json(下次自動跳過)。中途 Ctrl+C 可安全停止。
常見問題
Q:Token 過期了怎麼辦? 重跑 node oauth.js 更新 FB_PAGE_ACCESS_TOKEN(填了 FB_APP_SECRET 會自動換成 60 天長期 token)。
Q:可以同時處理多篇貼文嗎? 可以,但 fetch.js 每次會覆寫 pending.json,所以一次處理一篇最單純;或先抓多篇、把各篇的 replies.json 累加後一次 node reply.js(留言 ID 全域唯一)。
Q:想自己改 Claude 生成的回覆? 直接編輯 replies.json 再 node reply.js。
Q:replied.json 不見了會怎樣? 會把所有留言當成沒回過,最壞情況是有些被回兩次。
Claude 在執行這個 skill 時的工作守則
- 不要寫程式去呼叫付費 AI API — 生成回覆是 Claude 在對話裡直接做。
- 讀 pending.json → 寫 replies.json 是核心任務:
- 用留言者的語言回覆(中文回中文、英文回英文、日文回日文…)
- emoji_only 留言沒文字可判語言,改用 username 命名風格推測;判斷不出來就走英語安全牌,不亂猜被誤射
- 配合
post_caption的主題與語氣;友善、簡短(1~2 句)、像真人,不要罐頭 - 避免承諾、避免敏感話題、避免提到價格細節
- 每則回覆可含輕量 CTA(追蹤/分享/Tag 朋友/預告下一篇),類型在同一批裡輪流變化;約 70~80% 有 CTA 即可,留一些純情感回應避免整體像行銷帳號。
- 同一人連發多則留言 → 每則都回,但內容、emoji、CTA 都要彼此不同(對方是真愛粉,只回一條反而冷淡)。
- 寫完 replies.json 後,提醒/協助使用者執行
node reply.js送出。 - 若 token 失效,清楚告訴使用者重跑
node oauth.js。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

