browser-workspace 0.4.0

package/platforms/xhs/actions/_manage.py

@@ -0,0 +1,355 @@
+import json, time, re
+
+CFG = json.load(open("__CFG_PATH__"))
+OP = CFG["operation"]
+TARGET_NOTE_ID = CFG.get("note_id")
+TARGET_TITLE = CFG.get("title")
+
+
+
+def open_or_reuse_xhs(url):
+    current=current_tab()
+    current_url=current.get("url","")
+    if "xiaohongshu.com" in current_url:
+        target=current
+    else:
+        tabs=[t for t in list_tabs() if "xiaohongshu.com" in (t.get("url") or "")]
+        target=tabs[0] if tabs else None
+    if target:
+        switch_tab(target)
+        if page_info()["url"] != url:
+            goto_url(url)
+    else:
+        new_tab(url)
+    return current_tab()
+
+def out(payload):
+    print(json.dumps(payload, ensure_ascii=False, separators=(",", ":")))
+
+
+def resolve_manager_card():
+    open_or_reuse_xhs("https://creator.xiaohongshu.com/new/note-manager")
+    try:
+        wait_for_load()
+    except Exception:
+        pass
+    raw=""
+    for _ in range(20):
+        time.sleep(.5)
+        try:
+            raw = js('Array.from(document.querySelectorAll(".note-card")).map(function(c){var t=c.querySelector(".note-card__title"); return JSON.stringify({title:t?(t.innerText||t.textContent||"").trim():"", impression:c.dataset.impression||"", text:(c.innerText||"").trim()})}).join("\\n")') or ""
+        except Exception:
+            raw=""
+        if raw:
+            break
+    rows=[]
+    for line in (raw or "").splitlines():
+        try:
+            row=json.loads(line)
+            imp=json.loads(row.get("impression") or "{}")
+            row["note_id"]=(((imp.get("noteTarget") or {}).get("value") or {}).get("noteId"))
+            rows.append(row)
+        except Exception:
+            pass
+    for row in rows:
+        if TARGET_NOTE_ID and row.get("note_id")==TARGET_NOTE_ID:
+            return row
+        if TARGET_TITLE and row.get("title")==TARGET_TITLE:
+            return row
+    return None
+
+
+def card_exists_in_current_tab(note_id, title):
+    return js('Array.from(document.querySelectorAll(".note-card")).some(function(c){var t=c.querySelector(".note-card__title"); var ttl=t?(t.innerText||t.textContent||"").trim():""; var imp=c.dataset.impression||""; return '+json.dumps(bool(note_id))+' ? imp.indexOf('+json.dumps(note_id or "")+')>=0 : ttl==='+json.dumps(title or "")+'}) ? "yes" : "no"') == "yes"
+
+
+def resolve_status(row):
+    text=row.get("text") or ""
+    if "审核中" in text:
+        return "reviewing","审核中"
+    if "未通过" in text:
+        return "rejected","未通过"
+    for label,status in [("审核中","reviewing"),("未通过","rejected"),("已发布","published")]:
+        clicked=js('var es=Array.from(document.querySelectorAll("*")).filter(function(e){return e.children.length===0 && e.offsetParent && (e.innerText||e.textContent||"").trim()==='+json.dumps(label)+'}); if(es.length){es[0].click();"yes"}else{"no"}')
+        if clicked=="yes":
+            time.sleep(1.2)
+            if card_exists_in_current_tab(row.get("note_id"), row.get("title")):
+                return status,label
+    return "unknown","unknown"
+
+
+def open_update_editor(row):
+    note_id=row.get("note_id") or ""
+    title=row.get("title") or ""
+    clicked=js('var cards=Array.from(document.querySelectorAll(".note-card")); var c=cards.find(function(c){var t=c.querySelector(".note-card__title"); var ttl=t?(t.innerText||t.textContent||"").trim():""; var imp=c.dataset.impression||""; return '+json.dumps(bool(note_id))+' ? imp.indexOf('+json.dumps(note_id)+')>=0 : ttl==='+json.dumps(title)+'}); if(!c){"missing-card"}else{var a=c.querySelectorAll(".note-card__action-btn")[3]; if(a){a.click();"clicked"}else{"missing-edit"}}')
+    if clicked!="clicked":
+        return None, clicked
+    for _ in range(30):
+        time.sleep(.5)
+        url=page_info()["url"]
+        if "/publish/update" in url:
+            tab=current_tab()
+            return {"update_url":url,"target_id":tab.get("targetId") or tab.get("target_id")}, None
+    return None,"update_editor_timeout"
+
+
+def editor_media_kind():
+    accepts=""
+    for _ in range(30):
+        try:
+            accepts=js('Array.from(document.querySelectorAll("input[type=file]")).map(function(e){return e.accept||""}).join("|")') or ""
+        except Exception:
+            accepts=""
+        if accepts:
+            break
+        time.sleep(.5)
+    if any(ext in accepts for ext in (".mp4",".mov",".mkv",".mpeg")):
+        return "video"
+    if any(ext in accepts for ext in (".jpg",".jpeg",".png",".webp")):
+        return "image"
+    return "unknown"
+
+
+def set_update_fields(editor):
+    target_id=editor.get("target_id")
+    if target_id:
+        switch_tab(target_id)
+    if CFG.get("video"):
+        if editor_media_kind()!="video":
+            return False,"media_type_mismatch_video"
+        upload_file('input[type=file][accept*=".mp4"],input.upload-input', CFG["video"])
+        if target_id:
+            switch_tab(target_id)
+        for _ in range(120):
+            time.sleep(1)
+            text=js('(document.body.innerText||"").slice(0,1800)') or ""
+            # Once the editor is usable again and no obvious upload-progress copy is visible,
+            # proceed; final publish button enabled state is checked below.
+            if "上传中" not in text and "处理中" not in text:
+                break
+    if CFG.get("images"):
+        if editor_media_kind()!="image":
+            return False,"media_type_mismatch_image"
+        selector='input[type=file][accept*=".jpg"],input[type=file][accept*=".jpeg"],input[type=file][accept*=".png"]'
+        # upload_file accepts one file per call; replacing one image is
+        # enough for current Vlog use. Multiple-image update remains explicit and bounded.
+        first=True
+        for image in CFG["images"]:
+            upload_file(selector, image)
+            if target_id:
+                switch_tab(target_id)
+            time.sleep(1.5)
+            first=False
+    if target_id:
+        switch_tab(target_id)
+    if CFG.get("new_title") is not None:
+        fill_input('input[placeholder="填写标题会有更多赞哦"]', CFG["new_title"], clear_first=True, timeout=20)
+    if CFG.get("body") is not None:
+        fill_input('.tiptap.ProseMirror,[contenteditable=true]', CFG["body"], clear_first=True, timeout=20)
+    return True,None
+
+
+def submit_update():
+    # Upload processing can keep the publish control disabled for a while. Wait for
+    # the real control to become enabled rather than failing immediately.
+    clicked=False
+    for _ in range(180):
+        try:
+            state=js('var b=Array.from(document.querySelectorAll("button,[role=button]")).find(function(e){return e.offsetParent && (e.innerText||e.textContent||"").trim()==="发布"}); if(!b){"missing"}else if(b.disabled || (b.getAttribute("aria-disabled")||"false")==="true"){"disabled"}else{b.click();"clicked"}')
+            if state=="clicked":
+                clicked=True
+                break
+            if state=="missing":
+                nodes=cdp("Accessibility.getFullAXTree")["nodes"]
+                target=next((n for n in nodes if (n.get("role") or {}).get("value")=="button" and (n.get("name") or {}).get("value")=="发布" and n.get("backendDOMNodeId")),None)
+                if target:
+                    obj=cdp("DOM.resolveNode", backendNodeId=target["backendDOMNodeId"])["object"]["objectId"]
+                    value=cdp("Runtime.callFunctionOn", objectId=obj, functionDeclaration='function(){if(this.disabled||this.getAttribute("aria-disabled")==="true")return "disabled";this.click();return "clicked"}', returnByValue=True)["result"].get("value")
+                    if value=="clicked":
+                        clicked=True
+                        break
+        except Exception:
+            pass
+        time.sleep(1)
+    if not clicked:
+        return False,"publish_button_missing_or_disabled"
+    for _ in range(30):
+        time.sleep(1)
+        try:
+            url=js('location.href') or ""
+            if "/publish/success" in url or "published=true" in url:
+                return True,None
+        except Exception:
+            pass
+    return False,"update_submit_unverified"
+
+
+def own_profile_url():
+    open_or_reuse_xhs("https://www.xiaohongshu.com")
+    time.sleep(3)
+    href=js('var a=Array.from(document.querySelectorAll("a[href*=\\"/user/profile/\\"]")).find(function(a){return (a.innerText||a.textContent||"").trim()==="我"}); a?a.href:""')
+    return href or None
+
+
+def open_post_detail(note_id):
+    profile=own_profile_url()
+    if not profile:
+        return None, "own_profile_not_found"
+    goto_url(profile); time.sleep(3)
+    selector='section.note-item[data-note-id="'+note_id+'"]'
+    href=""
+    for _ in range(30):
+        href=js('var e=document.querySelector('+json.dumps(selector)+' ); var a=e&&e.querySelector("a.cover,a.title"); a?a.href:""')
+        if href:
+            break
+        js('window.scrollBy(0, Math.max(window.innerHeight*0.85, 600)); "scrolled"')
+        time.sleep(.6)
+    if not href:
+        return None, "post_not_visible_in_my_posts"
+    # The My Posts card carries the live xsec token. Following that route opens the
+    # post detail overlay/page; raw /explore/<id> URLs can 404 without this token.
+    goto_url(href); time.sleep(3)
+    ready=js('document.querySelector(".note-container") ? "yes" : "no"')
+    if ready!="yes":
+        return None, "post_detail_not_opened"
+    return {"profile_url":profile,"detail_url":page_info()["url"]}, None
+
+
+def expand_replies():
+    for _ in range(8):
+        n=js('var es=Array.from(document.querySelectorAll(".comments-el *")).filter(function(e){var t=(e.innerText||e.textContent||"").trim(); return e.offsetParent && e.children.length===0 && (/^展开\\s*\\d+\\s*条回复$/.test(t) || t==="展开更多回复" || t==="展开更多")}); es.slice(0,20).forEach(function(e){e.click()}); es.length')
+        try:
+            count=int(n or 0)
+        except Exception:
+            count=0
+        if count<=0:
+            break
+        time.sleep(.8)
+
+
+def detail_comments(limit):
+    expand_replies()
+    raw=js('Array.from(document.querySelectorAll(".comment-item")).map(function(c,i){var author=c.querySelector(".author .name"); var content=c.querySelector(".content .note-text"); var date=c.querySelector(".date > span:first-child"); var loc=c.querySelector(".date .location"); var reply=c.querySelector(".reply .count"); return JSON.stringify({dom_index:i,comment_id:(c.id||"").replace(/^comment-/,""),author:author?(author.innerText||author.textContent||"").trim():"",text:content?(content.innerText||content.textContent||"").trim():"",date:date?(date.innerText||date.textContent||"").trim():"",location:loc?(loc.innerText||loc.textContent||"").trim():"",is_reply:c.classList.contains("comment-item-sub"),reply_label:reply?(reply.innerText||reply.textContent||"").trim():"",replyable:!!c.querySelector(".reply")})}).join("\\n")')
+    rows=[]
+    for line in (raw or "").splitlines():
+        try:
+            row=json.loads(line)
+            row["comment_index"]=len(rows)
+            rows.append(row)
+            if len(rows)>=limit:
+                break
+        except Exception:
+            pass
+    return rows
+
+
+def activate_composer_for_comment():
+    # Top-level composer is present but inert until the visible "评论" affordance is clicked.
+    active=js('var e=document.querySelector("p.content-input[contenteditable=true]"); e&&e.offsetParent ? "yes" : "no"')
+    if active=="yes":
+        return True
+    clicked=js('var e=document.querySelector(".content-edit .inner-when-not-active, .content-edit .inner"); if(e){e.click();"yes"}else{"no"}')
+    if clicked!="yes":
+        return False
+    time.sleep(.4)
+    return js('var e=document.querySelector("p.content-input[contenteditable=true]"); e&&e.offsetParent ? "yes" : "no"')=="yes"
+
+
+def type_and_send(text):
+    focused=js('var e=document.querySelector("p.content-input[contenteditable=true]"); if(e&&e.offsetParent){e.focus();e.click();"yes"}else{"no"}')
+    if focused!="yes":
+        return False,"comment_editor_missing"
+    type_text(text)
+    time.sleep(.2)
+    sent=js('var b=document.querySelector("button.btn.submit,button.submit"); if(b && b.offsetParent && !b.disabled){b.click();"yes"}else{"no"}')
+    if sent!="yes":
+        return False,"comment_submit_missing_or_disabled"
+    time.sleep(1.5)
+    return True,None
+
+
+row=resolve_manager_card()
+if not row:
+    out({"ok":False,"operation":OP,"error":"post_not_found","note_id":TARGET_NOTE_ID,"title":TARGET_TITLE})
+    raise SystemExit(2)
+
+if OP=="status":
+    status,label=resolve_status(row)
+    out({"ok":True,"operation":"status","note_id":row.get("note_id"),"title":row.get("title"),"status":status,"status_label":label,"card_text":row.get("text")})
+    raise SystemExit(0)
+
+if OP=="update":
+    editor,err=open_update_editor(row)
+    if err:
+        out({"ok":False,"operation":"update","error":err,"note_id":row.get("note_id"),"title":row.get("title")}); raise SystemExit(5)
+    ok,err=set_update_fields(editor)
+    if not ok:
+        out({"ok":False,"operation":"update","error":err,"note_id":row.get("note_id"),"title":row.get("title"),"media_kind":editor_media_kind()}); raise SystemExit(4)
+    submit_ok,submit_err=submit_update()
+    # XHS edits sometimes apply without a stable success redirect. Platform-side
+    # note-manager verification is authoritative and prevents unsafe blind retries.
+    open_or_reuse_xhs("https://creator.xiaohongshu.com/new/note-manager")
+    try:
+        wait_for_load()
+    except Exception:
+        pass
+    time.sleep(2)
+    new_title=CFG.get("new_title") or row.get("title")
+    found=card_exists_in_current_tab(row.get("note_id"), new_title)
+    if found:
+        out({"ok":True,"operation":"update","note_id":row.get("note_id"),"old_title":row.get("title"),"title":new_title,"media_replaced":bool(CFG.get("video") or CFG.get("images")),"verified":True,"submit_redirect_verified":submit_ok})
+        raise SystemExit(0)
+    out({"ok":False,"operation":"update","error":submit_err or "platform_verification_failed","note_id":row.get("note_id"),"old_title":row.get("title"),"title":new_title,"verified":False})
+    raise SystemExit(8)
+
+route,err=open_post_detail(row.get("note_id"))
+if err:
+    out({"ok":False,"operation":OP,"error":err,"note_id":row.get("note_id"),"title":row.get("title")})
+    raise SystemExit(6)
+
+if OP=="comments":
+    rows=detail_comments(int(CFG.get("limit",50)))
+    out({"ok":True,"operation":"comments","note_id":row.get("note_id"),"title":row.get("title"),"source":"my_posts_detail","detail_url":route["detail_url"],"count":len(rows),"comments":rows})
+    raise SystemExit(0)
+
+if OP=="comment":
+    if not activate_composer_for_comment():
+        out({"ok":False,"operation":"comment","error":"comment_editor_missing"}); raise SystemExit(5)
+    ok,send_err=type_and_send(CFG["text"])
+    if not ok:
+        out({"ok":False,"operation":"comment","error":send_err}); raise SystemExit(5)
+    verified=js('Array.from(document.querySelectorAll(".comment-item .content .note-text")).some(function(e){return (e.innerText||e.textContent||"").trim()==='+json.dumps(CFG["text"])+ '}) ? "yes" : "no"')=="yes"
+    out({"ok":verified,"operation":"comment","note_id":row.get("note_id"),"title":row.get("title"),"text":CFG["text"],"verified":verified})
+    raise SystemExit(0 if verified else 7)
+
+if OP=="reply":
+    rows=detail_comments(200)
+    target=None
+    if CFG.get("comment_id"):
+        target=next((r for r in rows if r.get("comment_id")==CFG["comment_id"]),None)
+    elif CFG.get("comment_index") is not None:
+        idx=int(CFG["comment_index"]); target=next((r for r in rows if r["comment_index"]==idx),None)
+    else:
+        needle=CFG.get("contains") or ""
+        matches=[r for r in rows if needle in (r.get("text") or "")]
+        if len(matches)==1: target=matches[0]
+        elif len(matches)>1:
+            out({"ok":False,"operation":"reply","error":"ambiguous_comment_match","matches":len(matches)}); raise SystemExit(3)
+    if not target:
+        out({"ok":False,"operation":"reply","error":"comment_not_found"}); raise SystemExit(2)
+    if not target.get("replyable"):
+        out({"ok":False,"operation":"reply","error":"comment_not_replyable","comment":target}); raise SystemExit(4)
+    cid=target.get("comment_id")
+    opened=js('var c=document.getElementById('+json.dumps("comment-"+(cid or ""))+'); var r=c&&c.querySelector(".reply"); if(r){r.click();"yes"}else{"no"}')
+    if opened!="yes":
+        out({"ok":False,"operation":"reply","error":"reply_ui_missing"}); raise SystemExit(5)
+    time.sleep(.4)
+    ok,send_err=type_and_send(CFG["text"])
+    if not ok:
+        out({"ok":False,"operation":"reply","error":send_err}); raise SystemExit(5)
+    verified=js('Array.from(document.querySelectorAll(".comment-item .content .note-text")).some(function(e){return (e.innerText||e.textContent||"").trim()==='+json.dumps(CFG["text"])+ '}) ? "yes" : "no"')=="yes"
+    out({"ok":verified,"operation":"reply","note_id":row.get("note_id"),"title":row.get("title"),"replied_to":target,"text":CFG["text"],"verified":verified})
+    raise SystemExit(0 if verified else 7)
+
+out({"ok":False,"error":"unsupported_operation","operation":OP})
+raise SystemExit(2)

package/platforms/xhs/actions/_post.py

@@ -0,0 +1,192 @@
+import time, json, os
+
+CFG = json.load(open("__CFG_PATH__"))
+
+# NOTE: XHS overrides JSON.stringify (anti-bot). This script avoids it entirely.
+# Use js() for scalar returns only; use .click() instead of click_at_xy with rects.
+
+mode = CFG.get("mode", "image")
+
+def open_or_reuse_xhs(url):
+    current=current_tab()
+    current_url=current.get("url","")
+    if "xiaohongshu.com" in current_url:
+        target=current
+    else:
+        tabs=[t for t in list_tabs() if "xiaohongshu.com" in (t.get("url") or "")]
+        target=tabs[0] if tabs else None
+    if target:
+        switch_tab(target)
+        if page_info()["url"] != url:
+            goto_url(url)
+    else:
+        new_tab(url)
+    return current_tab()
+
+images = CFG.get("images", [CFG.get("image", "")])
+video = CFG.get("video", "")
+tags = CFG.get("tags", [])
+
+# Calculate steps dynamically
+base_steps = 6  # navigate, tab, upload, title, body, action
+extra = 0
+if mode == "image" and len(images) > 1:
+    extra += 1
+if mode == "video":
+    extra += 1  # wait for upload processing
+if tags:
+    extra += 1
+total_steps = base_steps + extra
+step = [0]
+def next_step(label):
+    step[0] += 1
+    print(f"[{step[0]}/{total_steps}] {label}")
+
+next_step("Navigating to XHS creator...")
+open_or_reuse_xhs("https://creator.xiaohongshu.com/publish/publish")
+wait_for_load()
+time.sleep(3)
+
+url = page_info()["url"]
+if "login" in url:
+    print("ERROR: Not logged in. Log in at creator.xiaohongshu.com first.")
+    capture_screenshot()
+    raise SystemExit(1)
+
+if mode == "video":
+    # Video tab is the default (first) tab — no need to click
+    next_step("Using video upload tab (default)...")
+    time.sleep(1)
+
+    next_step("Uploading video: " + video)
+    # Wait for the upload input; XHS can render the tab shell before mounting the file input.
+    for _ in range(20):
+        ready = js('document.querySelector(".upload-input,input[type=file]") ? "ready" : "wait"')
+        if ready == "ready":
+            break
+        time.sleep(0.5)
+    # Try upload_file first, fall back to DOM.performSearch for hidden inputs
+    try:
+        upload_file(".upload-input,input[type=file][accept*=video],input[type=file]", video)
+    except Exception:
+        doc = cdp("DOM.getDocument", depth=-1, pierce=True)
+        sr = cdp("DOM.performSearch", query='input[type="file"]')
+        if sr.get("resultCount", 0) > 0:
+            nodes = cdp("DOM.getSearchResults", searchId=sr["searchId"], fromIndex=0, toIndex=1)
+            nid = nodes["nodeIds"][0]
+            cdp("DOM.setFileInputFiles", files=[video], nodeId=nid)
+            cdp("DOM.discardSearchResults", searchId=sr["searchId"])
+        else:
+            print("ERROR: Cannot find file input for video upload.")
+            capture_screenshot()
+            raise SystemExit(1)
+
+    next_step("Waiting for video upload...")
+    for attempt in range(120):
+        time.sleep(5)
+        # Check if upload is complete by looking for video preview or progress
+        done = js('document.querySelector("[class*=video-player],[class*=videoPreview],video,.upload-success,[class*=uploadDone]") ? "ready" : "wait"')
+        if done == "ready":
+            print("  -> Upload done!")
+            break
+        if attempt % 6 == 5:
+            pct = js('var el=document.querySelector("[class*=progress],.percent,[class*=uploadProgress]"); el ? el.textContent : "..."')
+            print(f"  -> Uploading... {pct} ({(attempt+1)*5}s)")
+    else:
+        print("WARNING: Upload may still be processing.")
+        capture_screenshot()
+    time.sleep(2)
+
+else:
+    # Image mode — click the 2nd tab (上传图文)
+    next_step("Switching to image+text tab...")
+    js('var tabs=document.querySelectorAll(".creator-tab-item,.tab-item,[class*=publishTypeTabs] span,[class*=publish-type] span");if(tabs.length>=2){tabs[1].click()}else{var all=document.querySelectorAll("span");for(var i=0;i<all.length;i++){if(all[i].textContent.indexOf("\\u56fe\\u6587")>=0){all[i].click();break}}};"done"')
+    time.sleep(2)
+
+    check = js('document.querySelector(".upload-input") ? "found" : "notfound"')
+    if check != "found":
+        js('var spans=document.querySelectorAll("span,div,a");for(var i=0;i<spans.length;i++){var t=spans[i].textContent;if(t&&t.indexOf("\\u56fe\\u6587")>=0&&spans[i].offsetWidth>0){spans[i].click();break}};"retry"')
+        time.sleep(2)
+
+    next_step("Uploading image: " + images[0])
+    upload_file(".upload-input", images[0])
+    time.sleep(4)
+
+    # Upload additional images if any
+    if len(images) > 1:
+        next_step(f"Uploading {len(images)-1} more image(s)...")
+        for i, img in enumerate(images[1:], 2):
+            add_btn = js('document.querySelector(".add-upload,.upload-more,.add-btn,[class*=addImage]") ? "found" : "notfound"')
+            if add_btn == "found":
+                upload_file(".add-upload,.upload-more,.add-btn,[class*=addImage] input[type=file],.upload-input", img)
+            else:
+                upload_file(".upload-input", img)
+            print(f"  -> Image {i}/{len(images)}: {os.path.basename(img)}")
+            time.sleep(3)
+
+next_step("Setting title: " + CFG["title"])
+title_escaped = json.dumps(CFG["title"])
+title_set = js("var inp=document.querySelector('input[placeholder=\"填写标题会有更多赞哦\"],.c-input_inner input,input[placeholder]');if(inp){var ns=Object.getOwnPropertyDescriptor(HTMLInputElement.prototype,'value').set;ns.call(inp," + title_escaped + " );inp.dispatchEvent(new Event('input',{bubbles:true}));inp.dispatchEvent(new Event('change',{bubbles:true}));'done'}else{'missing'}")
+if title_set != "done":
+    print("ERROR: title input not found")
+    capture_screenshot()
+    raise SystemExit(1)
+time.sleep(0.5)
+
+next_step("Setting body text...")
+# Click into the editor area then type
+editor_set = js('var ed=document.querySelector(".tiptap.ProseMirror,#post-textarea,.ql-editor,[contenteditable=true]");if(ed){ed.focus();ed.click();"done"}else{"missing"}')
+if editor_set != "done":
+    print("ERROR: body editor not found")
+    capture_screenshot()
+    raise SystemExit(1)
+time.sleep(0.3)
+
+# Build body with tags appended as hashtags
+body = CFG["body"]
+if tags:
+    tag_str = " " + " ".join(f"#{t}" for t in tags)
+    body = body + tag_str
+type_text(body)
+time.sleep(0.5)
+
+if tags:
+    next_step(f"Tags appended: {', '.join(tags)}")
+
+if CFG["action"] == "publish":
+    next_step("Publishing...")
+    publish_result = js('var btn=[...document.querySelectorAll("button,[role=button]")].find(function(b){return b.offsetParent && (b.innerText||b.textContent||"").trim()==="发布"}) || document.querySelector(".publishBtn,.publish-btn,button.css-k01wfk,[class*=submit]");if(btn){btn.scrollIntoView({block:"center"});btn.click();"clicked"}else{"missing"}')
+    if publish_result != "clicked":
+        # XHS may expose the publish control only through the accessibility tree.
+        nodes = cdp("Accessibility.getFullAXTree")["nodes"]
+        target = next((n for n in nodes if (n.get("role") or {}).get("value") == "button" and (n.get("name") or {}).get("value") == "发布" and n.get("backendDOMNodeId")), None)
+        if target:
+            box = cdp("DOM.getBoxModel", backendNodeId=target["backendDOMNodeId"])["model"]["content"]
+            x = sum(box[0::2]) / 4
+            y = sum(box[1::2]) / 4
+            click_at_xy(x, y)
+            publish_result = "clicked"
+        else:
+            print("ERROR: publish button not found")
+            capture_screenshot()
+            raise SystemExit(1)
+    verified = False
+    for _ in range(20):
+        time.sleep(1)
+        state = js('location.href.indexOf("publish/success")>=0 || /发布成功|成功发布/.test(document.body.innerText||"") ? "success" : "wait"')
+        if state == "success":
+            verified = True
+            break
+    if not verified:
+        print("WARNING: publish click completed but success state is not yet verified")
+else:
+    next_step("Saving as draft...")
+    # Find 暂存 button by unicode
+    js('var btns=document.querySelectorAll("button");for(var i=0;i<btns.length;i++){if(btns[i].textContent.indexOf("\\u6682\\u5b58")>=0){btns[i].click();break}};"done"')
+
+time.sleep(3)
+capture_screenshot()
+if CFG["action"] == "publish":
+    print("Done! Post published." if verified else "Done! Post submitted; verification pending.")
+else:
+    print("Done! Post saved as draft.")

package/platforms/xhs/manifest.yaml

@@ -0,0 +1,98 @@
+platform: xhs
+status: migrated_unverified
+last_verified: null
+actions:
+  post: actions/_post.py
+  manage: actions/_manage.py
+verification:
+  status: migrated_unverified
+  last_verified: null
+  evidence: Existing post and creator-manager mechanics migrated from the working
+    post adapter. Current Browser Workspace live verification pending.
+url_patterns:
+- https://creator.xiaohongshu.com/*
+authentication:
+  required: true
+  notes: Use the caller's authenticated browser context; never store credentials.
+provenance:
+  source: skills/post/scripts/platforms/xhs
+  migrated_code: true
+flows:
+  post:
+    status: migrated_unverified
+    evidence: Migrated action; current Browser Workspace live verification pending.
+  manage:
+    status: migrated_unverified
+    evidence: Migrated manager action; current Browser Workspace live verification
+      pending.
+display_name: Xiaohongshu
+aliases:
+- xiaohongshu
+media:
+  modes:
+  - video
+  - images
+  video:
+    required_when_mode: video
+    formats:
+    - mp4
+  images:
+    required_when_mode: images
+    formats:
+    - jpg
+    - png
+    - webp
+    max_count: 18
+  mutually_exclusive:
+  - video
+  - images
+fields:
+  title:
+    cli: --title
+    required: true
+    type: string
+    max_chars: 20
+    implemented: true
+  body:
+    cli: --body
+    required: true
+    type: string
+    max_chars: 1000
+    implemented: true
+  tags:
+    cli: --tags
+    required: false
+    type: string_list
+    max_count: 5
+    implemented: true
+    render: 'append as #tag'
+  publish:
+    cli: --publish
+    required: false
+    type: boolean
+    default: false
+    implemented: true
+capabilities:
+  draft: true
+  publish: true
+  visibility: false
+  custom_thumbnail: false
+  series_or_collection: false
+  post_status: true
+  read_comments: true
+  write_comment: true
+  reply_comments: true
+content_mapping:
+  headline: title
+  main_text: body
+  discovery_terms: tags
+notes:
+- Without --publish, the adapter saves a draft.
+- After --publish, the XHS adapter automatically verifies the created note in creator
+  note manager and reports published/reviewing/rejected state.
+- comments/comment/reply resolve the post from the signed-in My Posts profile list
+  and use that card's live xsec detail route.
+- The opened post detail view is the canonical place to read comments, write a top-level
+  comment, and reply to comments.
+- The current adapter does not expose the Neo Vlog collection/series field even though
+  a prior manual browser-harness session created and attached a collection.

package/platforms/youtube/README.md

@@ -0,0 +1,10 @@
+---
+name: youtube-platform
+description: "YouTube Studio Browser Workspace upload mechanics."
+---
+
+# YouTube
+
+Use the parent Browser Workspace skill contract first. The
+`actions/` directory contains the canonical migrated Studio upload
+mechanics; `skills/post` retains payload and publishing orchestration.