@acedatacloud/skills 2026.713.1 → 2026.713.3

package/AGENTS.md

@@ -29,6 +29,7 @@ Skills are located in the `skills/` directory (also mirrored to `.agents/skills/
 ### AI Chat & Tools
 - **ai-chat** — Unified LLM gateway — GPT, Claude, Gemini, Kimi, Grok (50+ models)
 - **google-search** — Search the web, images, news, maps, places, and videos via Google
+- **tgstat** — Discover and analyze public Telegram channels/groups (BYOC username; no login)
 - **face-transform** — Face analysis, beautification, age/gender transform, swap, cartoon
 - **short-url** — Create and manage short URLs
 - **onepage-pdf** — Convert an HTML page into one tall single-page PDF (local; no API token; needs Python + pymupdf + Chrome/Edge)

package/README.md

@@ -48,6 +48,7 @@ Compatible with **30+ AI coding agents** via the [agentskills.io](https://agents
 |-------|-------------|
 | [ai-chat](skills/ai-chat/) | Unified LLM gateway — GPT, Claude, Gemini, Kimi, Grok (50+ models) |
 | [google-search](skills/google-search/) | Search the web, images, news, maps, places, and videos via Google |
+| [tgstat](skills/tgstat/) | Discover and analyze public Telegram channels/groups using a connected username as the default target |
 | [face-transform](skills/face-transform/) | Face analysis, beautification, age/gender transform, swap, cartoon |
 | [short-url](skills/short-url/) | Create and manage short URLs |
 | [onepage-pdf](skills/onepage-pdf/) | Convert an HTML page into one tall single-page PDF — no pagination breaks (local, no token) |
@@ -75,6 +76,7 @@ These skills drive third-party connectors users wire up at [auth.acedata.cloud/u
 | [didi-ride](skills/didi-ride/) | Book DiDi rides, estimate fares, query/cancel orders, plan routes via the DiDi MCP | `didi` (BYOC) |
 | [wecom](skills/wecom/) | WeCom (企业微信) self-built app — contacts, app messages, WeDoc, schedules, meetings | `wecom` (BYOC) |
 | [tencent-docs](skills/tencent-docs/) | Create / read / list / search / manage Tencent Docs — docs, sheets, slides, mind maps, flowcharts | `tencentdocs` (BYOC) |
+| [tgstat](skills/tgstat/) | Public Telegram source discovery, rankings, and audience research | `tgstat` (BYOC username) |
 
 ## Prerequisites
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@acedatacloud/skills",
-  "version": "2026.713.1",
+  "version": "2026.713.3",
   "description": "Agent Skills for AceDataCloud AI services — music, image, video generation, LLM chat, web search. Compatible with Claude Code, GitHub Copilot, Gemini CLI, OpenAI Codex, and 30+ AI coding agents.",
   "keywords": [
     "agent-skills",

package/skills/tgstat/SKILL.md

@@ -1,110 +1,136 @@
 ---
 name: tgstat
-description: Search Telegram channels/chats and pull audience + engagement stats (subscribers, reach, ER/ERR) from TGStat via its Stat API. Use when the user wants to find Telegram channels to advertise in (esp. Russian-language / RU market), research competitor channels, or check a channel's reach/engagement before buying an ad placement.
+description: "Research public Telegram channels and groups with TGStat. Use when discovering communities by topic, comparing public audience/reach/activity, checking a known @username, or shortlisting ad and outreach sources. The connector supplies TGSTAT_USERNAME as the default profile. Read-only: does not log in, join groups, scrape members, or send messages."
 when_to_use: |
-  Trigger when the user wants to discover Telegram channels by keyword /
-  category / country / language, inspect a specific channel's subscriber
-  count, average post reach and engagement (ER / ERR) to judge ad value,
-  or check their TGStat API quota. This is a marketing-research /
-  ad-channel-selection tool — it does NOT post to Telegram (use the
-  `telegram` connector for reading/sending messages).
+  Use for Telegram source discovery, competitor research, ad-channel
+  selection, and public audience analysis. It can find channels/chats by
+  keyword, inspect known public usernames, compare visible metrics, and use
+  the connected Telegram username as the default target. Use the separate
+  telegram connector for reading or sending messages in the user's account.
 connections: [tgstat]
-allowed_tools: [Bash]
+allowed_tools: [Bash, web_search, web_fetch]
 license: Apache-2.0
 metadata:
   author: acedatacloud
-  version: "1.0"
+  version: "2.1"
 ---
 
-Call the **TGStat Stat API** with `curl + jq`. The user's token is in
-`$TGSTAT_TOKEN` and is passed as the `token` query parameter on every request.
-Base URL: `https://api.tgstat.ru`.
+# TGStat Research
 
-Responses are JSON shaped `{"status":"ok","response": ...}`. Errors come back as
-`{"status":"error","error":"<message>"}` — show the `error` verbatim. An invalid
-token or an inactive/expired API plan will surface here; tell the user to check
-their token / plan in the TGStat 个人中心 and re-connect the connector.
+Use [scripts/tgstat.py](./scripts/tgstat.py) for public TGStat research. It uses
+only the Python standard library. Resolve the script at the start of every Bash
+call because each call runs in a fresh shell:
 
-**Always confirm the token + remaining quota first** (`usage/stat` is free and
-does not count against tariff quota):
+```bash
+TGSTAT="$SKILL_DIR/scripts/tgstat.py"; [ -f "$TGSTAT" ] || TGSTAT=$(find /tmp -maxdepth 8 -path '*/skills/*/scripts/tgstat.py' 2>/dev/null | head -1)
+[ -f "$TGSTAT" ] || { echo "tgstat script not found (SKILL_DIR=$SKILL_DIR)" >&2; exit 1; }
+python3 "$TGSTAT" profile
+```
+
+The connector injects `$TGSTAT_USERNAME`. It is not a login credential; it is
+only the default public channel/group target. Never print the full environment
+or treat the username as proof that the user owns the Telegram account.
+
+## Discover Sources by Topic
+
+Run `search` first:
+
+```bash
+TGSTAT="$SKILL_DIR/scripts/tgstat.py"; [ -f "$TGSTAT" ] || TGSTAT=$(find /tmp -maxdepth 8 -path '*/skills/*/scripts/tgstat.py' 2>/dev/null | head -1)
+[ -f "$TGSTAT" ] || { echo "tgstat script not found (SKILL_DIR=$SKILL_DIR)" >&2; exit 1; }
+python3 "$TGSTAT" search "Claude API" --type all --language English --country US
+```
+
+In public mode the command returns `web_queries`. Call `web_search` once per
+query, then:
+
+1. Keep only public `tgstat.com` regional-host and `t.me` results.
+2. Extract public `@username` values and deduplicate case-insensitively.
+3. Prefer results whose title/snippet directly matches the topic.
+4. Verify each shortlist entry with `info` or `stat` before reporting it.
+5. Include the source URL and say discovery may be incomplete because public
+   search-engine indexes are not TGStat's full database.
+
+TGStat's own keyword-search result endpoint requires sign-in. Do not try to
+bypass it, replay private AJAX endpoints, or claim public mode searches the
+full TGStat index.
+
+`--language` and `--country` are search terms for the web index, not guaranteed
+TGStat database filters.
+
+## Browse Public Rankings
+
+Use rankings when the user wants large or active public sources without a
+precise keyword:
 
 ```bash
-curl -sS "https://api.tgstat.ru/usage/stat?token=$TGSTAT_TOKEN" \
-  | jq '.status, (.response[]? | {serviceKey, title, spentChannels, spentRequests, expiredAt})'
+TGSTAT="$SKILL_DIR/scripts/tgstat.py"; [ -f "$TGSTAT" ] || TGSTAT=$(find /tmp -maxdepth 8 -path '*/skills/*/scripts/tgstat.py' 2>/dev/null | head -1)
+[ -f "$TGSTAT" ] || { echo "tgstat script not found (SKILL_DIR=$SKILL_DIR)" >&2; exit 1; }
+python3 "$TGSTAT" rankings --type channel --limit 20
+python3 "$TGSTAT" rankings --type chat --limit 20
+python3 "$TGSTAT" rankings --type channel --query crypto --limit 10
 ```
 
-## Find channels to advertise in (the main workflow)
+Channel cards can expose subscribers, one-post reach, and citation index.
+Chat cards can expose participants, recent message count, and MAU. Metrics are
+a current public snapshot, not a historical series.
+
+Public TGStat pages can intermittently return an authentication or rate-limit
+interstitial. When `rankings` returns `status: unavailable`, try the emitted
+`web_fetch_url`, then run its `web_queries` with `web_search`. Label those
+results as web-index discoveries, not as an authoritative TGStat rank.
 
-`GET /channels/search` — at least one of `q` (keyword, min 3 chars) or
-`category` is required.
+## Inspect a Known Channel or Group
 
-Params: `q`, `category`, `country`, `language` (default `russian`),
-`peer_type` (`channel` | `chat` | `all`, default `channel`),
-`search_by_description` (`0`/`1`), `limit` (max 100).
+Accept a public `@username`, bare username, `t.me/<username>` link, or a TGStat
+entity URL. Omit the target to inspect `$TGSTAT_USERNAME` from the connector:
 
 ```bash
-# Russian-language AI/ChatGPT channels, biggest first.
-curl -sS "https://api.tgstat.ru/channels/search" \
-  --data-urlencode "token=$TGSTAT_TOKEN" \
-  --data-urlencode "q=нейросети" \
-  --data-urlencode "language=russian" \
-  --data-urlencode "peer_type=channel" \
-  --data-urlencode "limit=50" -G \
-  | jq '.response.items | sort_by(-.participants_count)
-        | .[] | {username, title, subs: .participants_count, ci_index, link}'
+TGSTAT="$SKILL_DIR/scripts/tgstat.py"; [ -f "$TGSTAT" ] || TGSTAT=$(find /tmp -maxdepth 8 -path '*/skills/*/scripts/tgstat.py' 2>/dev/null | head -1)
+[ -f "$TGSTAT" ] || { echo "tgstat script not found (SKILL_DIR=$SKILL_DIR)" >&2; exit 1; }
+python3 "$TGSTAT" info
+python3 "$TGSTAT" info @durov
+python3 "$TGSTAT" stat https://t.me/example_public_chat
 ```
 
-- Use `--data-urlencode ... -G` so Cyrillic / spaces in `q` are encoded correctly.
-- `ci_index` (индекс цитирования) is TGStat's citation/authority score — higher =
-  more reposted/mentioned elsewhere, a useful quality signal beyond raw subs.
-- Try several keywords (`ChatGPT`, `нейросети`, `AI`, `разработка`, `API`) and
-  merge results; TGStat matches title/username (add `search_by_description=1` to
-  also match the channel description).
+In public mode, `info`/`stat` resolves whether the target is a channel or chat,
+returns public metadata, and enriches it with ranking metrics when the entity
+appears in the current public ranking. Empty metrics mean TGStat did not expose
+them publicly; do not call that full statistics.
 
-## Judge a channel's ad value
+For ad selection, rank by relevant reach/activity rather than subscriber count
+alone. High subscribers with weak reach or chat MAU can indicate an inactive or
+inflated audience. Present metrics as evidence, not a guarantee of lead quality.
 
-`GET /channels/stat?channelId=<@username | t.me/username | tgstat id>` returns
-the numbers that actually matter for ad pricing:
+## Check the Connected Profile
 
 ```bash
-curl -sS "https://api.tgstat.ru/channels/stat?token=$TGSTAT_TOKEN&channelId=@durov" \
-  | jq '.response | {
-      subs: .participants_count,
-      avg_post_reach,          # средний охват публикации
-      adv_reach_24h: .adv_post_reach_24h,   # рекламный охват за 24ч — key for CPM
-      er_percent, err_percent, err24_percent,
-      daily_reach, ci_index, posts_count
-    }'
+TGSTAT="$SKILL_DIR/scripts/tgstat.py"; [ -f "$TGSTAT" ] || TGSTAT=$(find /tmp -maxdepth 8 -path '*/skills/*/scripts/tgstat.py' 2>/dev/null | head -1)
+[ -f "$TGSTAT" ] || { echo "tgstat script not found (SKILL_DIR=$SKILL_DIR)" >&2; exit 1; }
+python3 "$TGSTAT" profile
 ```
 
-- For **ad CPM estimation** use `adv_post_reach_24h` (average *advertising* reach
-  of a post over 24h), not raw subscriber count — subs are vanity, reach is what
-  the ad actually gets seen by.
-- Low `err_percent` / `err24_percent` relative to subs = inflated/dead audience →
-  skip it.
-- For a **chat** (not channel) the response instead has `dau`/`wau`/`mau` and
-  `messages_count_*` fields.
-
-`GET /channels/get?channelId=...` returns descriptive info (title, about,
-`category`, `country`, `language`, `participants_count`, `ci_index`) when you
-just need to identify/verify a channel rather than full stats.
-
-## Reference data
-
-`GET /database/categories?token=$TGSTAT_TOKEN` lists valid `category` values you
-can pass to `channels/search`. There are also `/database/countries` and
-`/database/languages`. (See the docs for the full parameter/response shape.)
-
-## Gotchas
-
-- **Plan gating:** `channels/search` needs a **Stat API tariff S or higher**;
-  `channels/stat` / `channels/get` work on all Stat tariffs. If a call returns an
-  access error, the user's plan doesn't cover that method.
-- **Quota:** each unique channel and each request counts against the monthly
-  tariff (visible via `usage/stat`). Don't loop over hundreds of channels
-  blindly — search, shortlist, then `stat` only the shortlist.
-- **`q` min length is 3 chars** → `{"error":"param q is too short"}`.
-- **Cyrillic:** always send `q` via `--data-urlencode` (or pre-URL-encode) so the
-  keyword isn't mangled.
-- TGStat is a Russian service; the API and billing are RU-side — availability and
-  payment are the account owner's responsibility.
+`profile` reports the normalized public username used by target-less `info` and
+`stat`. It does not verify ownership or authenticate to Telegram/TGStat.
+
+## Outreach Research Workflow
+
+For prospecting or partnership research:
+
+1. Search several narrow problem/role keywords, not only broad `AI` terms.
+2. Shortlist sources by relevance first, then audience/reach/activity.
+3. Verify each public source and retain evidence URLs.
+4. Label each source as channel, group/chat, or uncertain.
+5. Produce a review queue with suggested angle and reason for fit.
+6. Require human approval before any message is sent through another connector.
+
+## Safety and Limits
+
+- Read-only. This skill does not join channels/groups, import invite links,
+  send messages, add members, or download member lists.
+- Reject private invite links (`t.me/+...`, `joinchat/...`) and message links.
+- Do not automate unsolicited bulk outreach or evade Telegram/TGStat controls.
+- Public pages and HTML can change; if parsing fails, use `web_fetch` for a
+  single public page and report only values visible in that page.
+- Public web discovery is incomplete and regional TGStat pages may differ.
+- Treat `$TGSTAT_USERNAME` as public profile context, not authentication.

package/skills/tgstat/scripts/tgstat.py

@@ -0,0 +1,561 @@
+#!/usr/bin/env python3
+"""Public TGStat research CLI with a connected Telegram username."""
+
+from __future__ import annotations
+
+import argparse
+import json
+import os
+import re
+import sys
+import urllib.error
+import urllib.parse
+import urllib.request
+from html.parser import HTMLParser
+from typing import Dict, List, Optional, Tuple, Union
+
+DEFAULT_PUBLIC_BASE = "https://tgstat.com"
+USER_AGENT = (
+    "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
+    "AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126 Safari/537.36"
+)
+PUBLIC_USERNAME_RE = re.compile(r"[A-Za-z0-9_]{3,}")
+TGSTAT_HOST_RE = re.compile(r"^(?:[a-z]{2,3}\.)?tgstat\.com$", re.IGNORECASE)
+TGSTAT_INPUT_HOST_RE = re.compile(r"^(?:(?:[a-z]{2,3}\.)?tgstat\.com|tgstat\.ru)$", re.IGNORECASE)
+ENTITY_PATH_RE = re.compile(r"/(channel|chat)/(@[A-Za-z0-9_]{3,}|id\d+)/stat/?", re.IGNORECASE)
+
+
+class TGStatError(RuntimeError):
+    pass
+
+
+def _compact(value: str) -> str:
+    return re.sub(r"\s+", " ", value or "").strip()
+
+
+def _metric_key(label: str) -> str:
+    key = re.sub(r"[^a-z0-9]+", "_", label.lower()).strip("_")
+    return key or "value"
+
+
+def _metric_number(value: str) -> Optional[Union[int, float]]:
+    compact = _compact(value).replace(" ", "").replace(",", "")
+    match = re.fullmatch(r"(-?\d+(?:\.\d+)?)([kmb])?", compact, re.IGNORECASE)
+    if not match:
+        return None
+    number = float(match.group(1))
+    multiplier = {"k": 1_000, "m": 1_000_000, "b": 1_000_000_000}.get(
+        (match.group(2) or "").lower(), 1
+    )
+    result = number * multiplier
+    return int(result) if result.is_integer() else result
+
+
+def _entity_from_url(url: str) -> Optional[Tuple[str, str]]:
+    path = urllib.parse.unquote(urllib.parse.urlparse(url).path)
+    match = ENTITY_PATH_RE.fullmatch(path)
+    if not match:
+        return None
+    return match.group(1).lower(), match.group(2)
+
+
+def _safe_url(url: str) -> str:
+    parsed = urllib.parse.urlsplit(url)
+    return urllib.parse.urlunsplit((parsed.scheme, parsed.netloc, parsed.path, "", ""))
+
+
+def _safe_error_body(body: str) -> str:
+    return _compact(body)[:200]
+
+
+def _validate_request_url(url: str, initial_host: Optional[str] = None) -> None:
+    parsed = urllib.parse.urlsplit(url)
+    host = parsed.hostname or ""
+    if parsed.scheme != "https":
+        raise TGStatError("TGStat requests and redirects must use HTTPS")
+    allowed_from = initial_host or host
+    if TGSTAT_HOST_RE.fullmatch(allowed_from):
+        if not TGSTAT_HOST_RE.fullmatch(host):
+            raise TGStatError(f"TGStat redirected to an unexpected host: {host}")
+    else:
+        raise TGStatError(f"unsupported TGStat request host: {allowed_from}")
+
+
+class SafeRedirectHandler(urllib.request.HTTPRedirectHandler):
+    def __init__(self, initial_host: str) -> None:
+        super().__init__()
+        self.initial_host = initial_host
+
+    def redirect_request(self, req, fp, code, msg, headers, newurl):
+        _validate_request_url(newurl, self.initial_host)
+        return super().redirect_request(req, fp, code, msg, headers, newurl)
+
+
+def _open_url(request: urllib.request.Request, timeout: int):
+    initial_url = request.full_url
+    _validate_request_url(initial_url)
+    initial_host = urllib.parse.urlsplit(initial_url).hostname or ""
+    opener = urllib.request.build_opener(SafeRedirectHandler(initial_host))
+    return opener.open(request, timeout=timeout)
+
+
+class RankingParser(HTMLParser):
+    def __init__(self) -> None:
+        super().__init__(convert_charrefs=True)
+        self.items: List[dict] = []
+        self.current: Optional[dict] = None
+        self.card_div_depth = 0
+        self.capture: Optional[dict] = None
+        self.pending_metric: Optional[str] = None
+
+    def handle_starttag(self, tag: str, attrs: List[Tuple[str, Optional[str]]]) -> None:
+        attr = {key: value or "" for key, value in attrs}
+        classes = set(attr.get("class", "").split())
+        if tag == "div" and "peer-item-row" in classes and self.current is None:
+            self.current = {"metrics": {}, "metrics_display": {}}
+            self.card_div_depth = 1
+            return
+        if self.current is None:
+            return
+        if tag == "div":
+            self.card_div_depth += 1
+        if self.capture is not None:
+            self.capture["depth"] += 1
+            return
+
+        if tag == "a" and attr.get("href"):
+            entity = _entity_from_url(attr["href"])
+            hostname = urllib.parse.urlparse(attr["href"]).hostname or ""
+            if entity and TGSTAT_HOST_RE.fullmatch(hostname):
+                peer_type, identifier = entity
+                self.current.setdefault("type", peer_type)
+                self.current.setdefault("identifier", identifier)
+                self.current.setdefault("tgstat_url", attr["href"])
+
+        kind = ""
+        if tag == "div" and "ribbon" in classes:
+            kind = "rank"
+        elif tag == "div" and {"text-truncate", "font-16", "text-dark"}.issubset(classes):
+            if self.current.get("tgstat_url") and not self.current.get("title"):
+                kind = "title"
+        elif tag == "span" and {"border", "rounded", "bg-light"}.issubset(classes):
+            if self.current.get("tgstat_url") and not self.current.get("category"):
+                kind = "category"
+        elif tag == "h4":
+            kind = "metric_value"
+        elif tag == "div" and {"text-muted", "text-truncate"}.issubset(classes):
+            if self.pending_metric is not None:
+                kind = "metric_label"
+        if kind:
+            self.capture = {"tag": tag, "kind": kind, "depth": 1, "parts": []}
+
+    def handle_endtag(self, tag: str) -> None:
+        if self.current is None:
+            return
+        if self.capture is not None:
+            self.capture["depth"] -= 1
+            if self.capture["depth"] == 0:
+                self._finish_capture()
+        if tag == "div":
+            self.card_div_depth -= 1
+            if self.card_div_depth == 0:
+                self._finish_card()
+
+    def handle_data(self, data: str) -> None:
+        if self.capture is not None:
+            self.capture["parts"].append(data)
+
+    def _finish_capture(self) -> None:
+        assert self.current is not None and self.capture is not None
+        kind = self.capture["kind"]
+        value = _compact("".join(self.capture["parts"]))
+        self.capture = None
+        if not value:
+            return
+        if kind == "rank":
+            match = re.search(r"\d+", value)
+            if match:
+                self.current["rank"] = int(match.group())
+        elif kind in {"title", "category"}:
+            self.current[kind] = value
+        elif kind == "metric_value":
+            self.pending_metric = value
+        elif kind == "metric_label" and self.pending_metric is not None:
+            key = _metric_key(value)
+            display = self.pending_metric
+            self.current["metrics_display"][key] = display
+            number = _metric_number(display)
+            self.current["metrics"][key] = number if number is not None else display
+            self.pending_metric = None
+
+    def _finish_card(self) -> None:
+        assert self.current is not None
+        item = self.current
+        identifier = item.get("identifier", "")
+        if identifier.startswith("@"):
+            item["username"] = identifier
+            item["telegram_url"] = f"https://t.me/{identifier[1:]}"
+        else:
+            item["username"] = None
+            item["telegram_url"] = None
+        if item.get("tgstat_url") and item.get("title"):
+            self.items.append(item)
+        self.current = None
+        self.card_div_depth = 0
+        self.capture = None
+        self.pending_metric = None
+
+
+class DetailParser(HTMLParser):
+    def __init__(self) -> None:
+        super().__init__(convert_charrefs=True)
+        self.meta: Dict[str, str] = {}
+        self.capture: Optional[dict] = None
+        self.pending_metric: Optional[str] = None
+        self.metrics: Dict[str, Union[int, float, str]] = {}
+        self.metrics_display: Dict[str, str] = {}
+        self.text_parts: List[str] = []
+
+    def handle_starttag(self, tag: str, attrs: List[Tuple[str, Optional[str]]]) -> None:
+        attr = {key.lower(): value or "" for key, value in attrs}
+        if tag == "meta":
+            key = (attr.get("property") or attr.get("name") or "").lower()
+            if key in {"og:title", "og:description", "description"} and attr.get("content"):
+                self.meta[key] = _compact(attr["content"])
+        if self.capture is not None:
+            self.capture["depth"] += 1
+            return
+        classes = set(attr.get("class", "").split())
+        kind = ""
+        if tag == "title":
+            kind = "title"
+        elif tag == "h4":
+            kind = "metric_value"
+        elif tag == "div" and {"text-muted", "text-truncate"}.issubset(classes):
+            if self.pending_metric is not None:
+                kind = "metric_label"
+        if kind:
+            self.capture = {"kind": kind, "depth": 1, "parts": []}
+
+    def handle_endtag(self, tag: str) -> None:
+        if self.capture is None:
+            return
+        self.capture["depth"] -= 1
+        if self.capture["depth"] != 0:
+            return
+        kind = self.capture["kind"]
+        value = _compact("".join(self.capture["parts"]))
+        self.capture = None
+        if kind == "title" and value:
+            self.meta.setdefault("title", value)
+        elif kind == "metric_value" and value:
+            self.pending_metric = value
+        elif kind == "metric_label" and value and self.pending_metric is not None:
+            key = _metric_key(value)
+            self.metrics_display[key] = self.pending_metric
+            number = _metric_number(self.pending_metric)
+            self.metrics[key] = number if number is not None else self.pending_metric
+            self.pending_metric = None
+
+    def handle_data(self, data: str) -> None:
+        text = _compact(data)
+        if text:
+            self.text_parts.append(text)
+        if self.capture is not None:
+            self.capture["parts"].append(data)
+
+
+def parse_rankings_html(html: str) -> List[dict]:
+    parser = RankingParser()
+    parser.feed(html)
+    return parser.items
+
+
+def parse_detail_html(html: str, url: str) -> dict:
+    parser = DetailParser()
+    parser.feed(html)
+    text = " ".join(parser.text_parts)
+    entity = _entity_from_url(url)
+    title = parser.meta.get("og:title") or parser.meta.get("title")
+    recognized = bool(entity and parser.meta.get("og:title") and title and "tgstat" in title.casefold())
+    result = {
+        "status": (
+            "restricted"
+            if "authentication required" in text.lower()
+            else "ok" if recognized else "unrecognized"
+        ),
+        "type": entity[0] if entity else None,
+        "identifier": entity[1] if entity else None,
+        "title": title,
+        "description": parser.meta.get("og:description") or parser.meta.get("description"),
+        "metrics": parser.metrics,
+        "metrics_display": parser.metrics_display,
+        "tgstat_url": url,
+    }
+    identifier = result["identifier"]
+    result["telegram_url"] = f"https://t.me/{identifier[1:]}" if isinstance(identifier, str) and identifier.startswith("@") else None
+    return result
+
+
+def _json_out(payload: object) -> None:
+    print(json.dumps(payload, ensure_ascii=False, indent=2))
+
+
+def _request_with_url(
+    url: str, timeout: int, data: Optional[bytes] = None, headers: Optional[dict] = None
+) -> Tuple[str, str]:
+    request_headers = {
+        "User-Agent": USER_AGENT,
+        "Accept": "application/json, text/html;q=0.9, */*;q=0.8",
+        "Accept-Language": "en-US,en;q=0.9",
+        **(headers or {}),
+    }
+    request = urllib.request.Request(url, data=data, headers=request_headers)
+    try:
+        with _open_url(request, timeout) as response:
+            return response.read().decode("utf-8", "replace"), response.geturl()
+    except urllib.error.HTTPError as exc:
+        body = exc.read().decode("utf-8", "replace")
+        raise TGStatError(f"HTTP {exc.code} from {_safe_url(url)}: {_safe_error_body(body)}") from exc
+    except urllib.error.URLError as exc:
+        raise TGStatError(f"network error for {_safe_url(url)}: {_safe_error_body(str(exc.reason))}") from exc
+
+
+def _request(url: str, timeout: int, data: Optional[bytes] = None, headers: Optional[dict] = None) -> str:
+    return _request_with_url(url, timeout, data, headers)[0]
+
+
+def _public_base(value: str) -> str:
+    parsed = urllib.parse.urlparse(value if "://" in value else f"https://{value}")
+    if parsed.scheme != "https" or not parsed.hostname or not TGSTAT_HOST_RE.fullmatch(parsed.hostname):
+        raise TGStatError("public host must be tgstat.com or a regional *.tgstat.com host")
+    return f"https://{parsed.hostname}"
+
+
+def _normalize_target(target: str) -> Tuple[Optional[str], Optional[str], Optional[str]]:
+    target = target.strip()
+    if target.startswith("http://") or target.startswith("https://"):
+        parsed = urllib.parse.urlparse(target)
+        if parsed.scheme != "https":
+            raise TGStatError("target URL must use HTTPS")
+        if parsed.hostname and TGSTAT_INPUT_HOST_RE.fullmatch(parsed.hostname):
+            entity = _entity_from_url(target)
+            if not entity:
+                raise TGStatError("TGStat target must be a channel or chat statistics URL")
+            identifier = entity[1]
+            canonical_url = target
+            if parsed.hostname.casefold() == "tgstat.ru":
+                canonical_url = urllib.parse.urlunsplit(("https", "tgstat.com", parsed.path, "", ""))
+            return identifier, entity[0], canonical_url
+        if parsed.hostname in {"t.me", "www.t.me"}:
+            parts = [part for part in parsed.path.split("/") if part]
+            if len(parts) != 1 or not PUBLIC_USERNAME_RE.fullmatch(parts[0]):
+                raise TGStatError("t.me target must be a public username link, not an invite or message link")
+            return f"@{parts[0]}", None, None
+        raise TGStatError("target URL must be a tgstat.com entity or t.me link")
+    if target.startswith("@") and PUBLIC_USERNAME_RE.fullmatch(target[1:]):
+        return target, None, None
+    if re.fullmatch(r"\d+", target):
+        return f"id{target}", None, None
+    if re.fullmatch(r"id\d+", target, re.IGNORECASE):
+        return f"id{target[2:]}", None, None
+    if PUBLIC_USERNAME_RE.fullmatch(target):
+        return f"@{target}", None, None
+    raise TGStatError("target must be @username, a t.me link, TGStat entity URL, or id<number>")
+
+
+def _connected_username() -> str:
+    value = os.environ.get("TGSTAT_USERNAME", "").strip()
+    if not value:
+        raise TGStatError("TGSTAT_USERNAME is not configured; reconnect TGStat with a public Telegram username")
+    identifier, _, _ = _normalize_target(value)
+    if not identifier or not identifier.startswith("@"):
+        raise TGStatError("TGSTAT_USERNAME must be a public Telegram username such as @durov")
+    return identifier
+
+
+def _target_or_profile(target: str) -> str:
+    return target.strip() or _connected_username()
+
+
+def command_profile(args: argparse.Namespace) -> None:
+    username = _connected_username()
+    _json_out(
+        {
+            "mode": "public",
+            "username": username,
+            "telegram_url": f"https://t.me/{username[1:]}",
+            "verified": False,
+        }
+    )
+
+
+def _web_queries(query: str, peer_type: str, language: str, country: str) -> List[str]:
+    suffix = " ".join(part for part in (language, country) if part).strip()
+    quoted = f'"{query}"'
+    kinds = [peer_type] if peer_type != "all" else ["chat", "channel"]
+    queries = [f"site:tgstat.com/{kind}/ {quoted} {suffix}".strip() for kind in kinds]
+    queries.extend(f"site:t.me/ {quoted} {kind} {suffix}".strip() for kind in kinds)
+    return queries
+
+
+def command_search(args: argparse.Namespace) -> None:
+    query = args.query.strip()
+    if len(query) < 3:
+        raise TGStatError("query must contain at least 3 characters")
+    _json_out(
+        {
+            "mode": "public",
+            "query": query,
+            "requires_web_search": True,
+            "web_queries": _web_queries(query, args.type, args.language, args.country),
+            "instructions": (
+                "Run web_search for each query, keep only tgstat.com and t.me results, "
+                "deduplicate by username, then inspect shortlisted TGStat URLs with the info command."
+            ),
+            "limitations": "TGStat's own keyword-search results require sign-in; public web indexes may be incomplete.",
+        }
+    )
+
+
+def _ranking_fallback(url: str, peer_type: str, query: str, reason: str) -> None:
+    subject = query.strip() or ("Telegram channels" if peer_type == "channel" else "Telegram groups")
+    _json_out(
+        {
+            "mode": "public",
+            "status": "unavailable",
+            "source": url,
+            "reason": reason,
+            "requires_web_fetch": True,
+            "web_fetch_url": url,
+            "requires_web_search": True,
+            "web_queries": [
+                f'site:tgstat.com/{peer_type}/ "{subject}"',
+                f'site:t.me/ "{subject}"',
+            ],
+            "limitations": "Fallback results are web-index discoveries, not an authoritative TGStat ranking.",
+        }
+    )
+
+
+def command_rankings(args: argparse.Namespace) -> None:
+    base = _public_base(args.host)
+    plural = "channels" if args.type == "channel" else "chats"
+    url = f"{base}/ratings/{plural}"
+    try:
+        html = _request(url, args.timeout)
+    except TGStatError as exc:
+        _ranking_fallback(url, args.type, args.query, str(exc))
+        return
+    normalized_html = html.casefold()
+    if "authentication required" in normalized_html or "too many requests" in normalized_html:
+        _ranking_fallback(url, args.type, args.query, "TGStat returned an authentication or rate-limit interstitial")
+        return
+    items = parse_rankings_html(html)
+    if not items:
+        _ranking_fallback(url, args.type, args.query, "TGStat ranking HTML was empty or not recognized")
+        return
+    if args.query:
+        needle = args.query.casefold()
+        items = [
+            item
+            for item in items
+            if needle in " ".join(str(item.get(key) or "") for key in ("title", "username", "category")).casefold()
+        ]
+    _json_out({"mode": "public", "source": url, "count": min(len(items), args.limit), "items": items[: args.limit]})
+
+
+def _public_info(args: argparse.Namespace, target: str, peer_type: str) -> dict:
+    identifier, inferred_type, exact_url = _normalize_target(target)
+    types = [inferred_type] if inferred_type else ([peer_type] if peer_type != "auto" else ["channel", "chat"])
+    urls = [exact_url] if exact_url else [f"{_public_base(args.host)}/{kind}/{identifier}/stat" for kind in types]
+    errors: List[str] = []
+    for url in urls:
+        if not url:
+            continue
+        try:
+            html, final_url = _request_with_url(url, args.timeout)
+            final_host = urllib.parse.urlparse(final_url).hostname or ""
+            if not TGSTAT_HOST_RE.fullmatch(final_host):
+                raise TGStatError(f"TGStat redirected to an unexpected host: {final_host}")
+            result = parse_detail_html(html, final_url)
+        except TGStatError as exc:
+            errors.append(str(exc))
+            continue
+        if result["status"] == "ok" and result.get("title"):
+            if not result["metrics"] and result.get("type") in {"channel", "chat"}:
+                plural = "channels" if result["type"] == "channel" else "chats"
+                ranking_url = f"{_public_base(args.host)}/ratings/{plural}"
+                try:
+                    ranking_items = parse_rankings_html(_request(ranking_url, args.timeout))
+                except TGStatError:
+                    ranking_items = []
+                normalized = str(result.get("identifier") or "").casefold()
+                snapshot = next(
+                    (item for item in ranking_items if str(item.get("identifier") or "").casefold() == normalized),
+                    None,
+                )
+                if snapshot:
+                    result["metrics"] = snapshot["metrics"]
+                    result["metrics_display"] = snapshot["metrics_display"]
+                    result["ranking_snapshot"] = {
+                        "rank": snapshot.get("rank"),
+                        "category": snapshot.get("category"),
+                        "source": ranking_url,
+                    }
+            return result
+        errors.append(f"restricted or empty public page: {url}")
+    raise TGStatError("; ".join(errors) or "could not resolve target on public TGStat pages")
+
+
+def command_info(args: argparse.Namespace) -> None:
+    _json_out({"mode": "public", **_public_info(args, _target_or_profile(args.target), args.type)})
+
+
+def command_stat(args: argparse.Namespace) -> None:
+    _json_out(
+        {"mode": "public", "limited_metrics": True, **_public_info(args, _target_or_profile(args.target), args.type)}
+    )
+
+
+def build_parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--host", default=os.environ.get("TGSTAT_PUBLIC_HOST", DEFAULT_PUBLIC_BASE))
+    parser.add_argument("--timeout", type=int, default=30)
+    sub = parser.add_subparsers(dest="command", required=True)
+
+    sub.add_parser("profile").set_defaults(func=command_profile)
+
+    search = sub.add_parser("search")
+    search.add_argument("query")
+    search.add_argument("--type", choices=("channel", "chat", "all"), default="all")
+    search.add_argument("--language", default="")
+    search.add_argument("--country", default="")
+    search.add_argument("--limit", type=int, default=20)
+    search.set_defaults(func=command_search)
+
+    rankings = sub.add_parser("rankings")
+    rankings.add_argument("--type", choices=("channel", "chat"), default="channel")
+    rankings.add_argument("--query", default="", help="Filter the public top list locally")
+    rankings.add_argument("--limit", type=int, default=20)
+    rankings.set_defaults(func=command_rankings)
+
+    for name, func in (("info", command_info), ("stat", command_stat)):
+        command = sub.add_parser(name)
+        command.add_argument("target", nargs="?", default="")
+        command.add_argument("--type", choices=("auto", "channel", "chat"), default="auto")
+        command.set_defaults(func=func)
+    return parser
+
+
+def main() -> None:
+    args = build_parser().parse_args()
+    try:
+        args.func(args)
+    except TGStatError as exc:
+        _json_out({"error": str(exc)})
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()