Install
$ agentstack add skill-tzengyuxio-skills-sichuan-mandarin ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ● Filesystem access Used
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
Sichuan Mandarin — 四川話 / 川渝話
Transform text into Sichuan Mandarin (四川話) — the Southwestern Mandarin variety spoken across Sichuan and Chongqing in the 成渝片 zone. The result should read like a Chengdu 茶館 regular cracking jokes, a Chongqing 火鍋 boss yelling at the kitchen, or 張麻子 in《讓子彈飛》laying down the law — direct, salted with 江湖 humor, and unmistakably 川渝.
四川話 is NOT broken Putonghua, nor is it the cartoon "麻將-and-火鍋" caricature pinned on Sichuan in TV comedy. It is a real Mandarin variety with its own phonology, lexicon, and pragmatic particle stack. Most speakers can also read and write Putonghua; this skill targets the written representation of casual 川渝 speech — the way those phonetic and lexical features surface in text messages, novels, screenplay dialogue, and online posts.
Arguments
$0— (Optional) Path to a markdown or text file. If omitted, ask for inline
text or find the most recent text in conversation context.
Step 1: Load the Text
- If
$0is provided, read the file. - If no argument, find the most recent text in the conversation, or ask the user.
- The input can be Mandarin Chinese (rewrite as Sichuan Mandarin), **any
language (translate to Sichuan Mandarin), or English** (translate to Sichuan Mandarin).
Step 2: Apply Sichuan Mandarin Patterns
Apply the rules below systematically. Goal: every 2-3 sentences should carry at least one 川渝 marker — a signature word, a 莫/莫得 negation, a 嘛/噻/嘞/哈 particle, a 「兒」化 ending, or a 江湖-toned construction. Not every rule applies to every sentence — pick what fits naturally.
四川話 sits on a continuum from light flavor (a few signature words on an otherwise standard backbone) up to heavy 江湖 cadence with 川劇丑角 self-mockery. This skill targets that full range, deliberately avoiding stereotype-laden caricature (lazy / mahjong-addicted / hotpot-obsessed tropes).
Soul Clause — 川渝江湖氣 [C-class Hard, all intensities]
The output must carry the 川渝 江湖+幽默口語 voice. Concretely, every piece of meaningful Medium/Heavy output must contain:
- ≥3 signature lexical markers drawn from {巴適,安逸,莫得,搞快點,龜兒,
撇脫,擺龍門陣,牙尖,假打,鬼火冒,哦豁,板眼兒,瓜娃子}.
- 「兒」化 density above Putonghua — Sichuan 兒化 is restricted (only 4
韻尾) but pervasive in the slots it occupies: place names (成都兒,重慶兒), small objects (雞蛋兒,板凳兒,棍兒,缽兒), playful adjectives (好玩兒, 乖乖兒).
- ≥1 句末助詞 per ~2 sentences, drawn from {嘛,噻,嘞,哈,撒,喲,哦}.
- 江湖+自嘲 cadence — short blunt clauses, rhetorical 反問, the speaker
willing to throw 「老子」 / 「龜兒」 at themselves and the world. The line between mock-anger and humor stays thin.
If the output lacks the soul clause, it is not 四川話 — it is only Putonghua with a few words swapped.
Grammar Patterns(語法模式)
G1. 「莫 / 莫得」否定 [Hard, Light+]
四川話's most identifiable grammatical marker. 莫 = 沒 (negation of past / existence); 莫得 = 沒有 (don't have).
> Standard: 我沒去。 > > Sichuan: 我莫去。
> Standard: 沒有錢。 > > Sichuan: 莫得錢。
> Standard: 沒辦法。 > > Sichuan: 莫得辦法。 / 莫法。
「莫」also doubles as a soft prohibitive (= 別 / 不要):
> Standard: 別亂說。 > > Sichuan: 莫亂說嘛。
G2. 「兒」化 — 受限但頻繁 [Hard, Light+]
四川話 has only 4 兒化 韻尾 (vs. Beijing's 26), but uses them densely in specific slots. Add 兒 化 freely to:
| 類型 | 範例 | |---|---| | 地名 | 成都兒、重慶兒、府南河兒 | | 小物件 | 雞蛋兒、板凳兒、缽兒、棍兒、刀兒、碗兒 | | 親暱稱呼 | 妹兒、兒娃兒、男娃兒、女娃兒 | | 形容詞輕語感 | 好玩兒、乖乖兒、慢慢兒、悄悄兒 | | 量詞 | 一陣兒、一哈兒(一下兒)、一坨兒 |
Do not add 兒 to abstract nouns (思想兒 ✗) or actions (跑兒 ✗) — that is northern Mandarin habit, not 川渝.
G3. 「搞」萬用動詞 [Hard, Light+]
「搞」is the 川渝 all-purpose verb — covers do/make/handle/mess-with. Often combined with 快點/緊(趕緊)/啥子/這樣 for full sentences.
| 套式 | 意思 | |---|---| | 搞快點 | 趕緊、加速 | | 搞啥子 | 在做什麼? | | 搞不贏 | 來不及、忙不過來 | | 搞不來 | 不會做、做不來 | | 亂搞 | 胡來 | | 整 | 「搞」的近義詞,也常用 |
> Standard: 你在做什麼? > > Sichuan: 你搞啥子嘛?
> Standard: 快點啦。 > > Sichuan: 搞快點噻!
G4. 反問 — 「啥子 / 咋個 / 哪個」 [Hard, Light+]
四川話 question words differ from Putonghua and ride a sharp rhetorical tone.
| 川渝 | 普通話 | |---|---| | 啥子 | 什麼 | | 咋個 | 怎麼 / 怎麼辦 | | 啷個 | 怎麼(變體) | | 哪個 | 誰 / 哪一個 | | 啥子時候 | 什麼時候 | | 咋的嘛 | 怎麼樣?/敢怎樣? |
> Standard: 你說什麼? > > Sichuan: 你說啥子?
> Standard: 那怎麼辦? > > Sichuan: 那咋個辦嘛?
G5. 動詞 + 「起 / 倒 / 起來」補語 [Soft, Medium+]
「起」「倒」 used as resultative or aspectual particles, slightly differently from Putonghua.
> Standard: 拿著。 > > Sichuan: 拿起。
> Standard: 站著。 > > Sichuan: 站起。 / 杵到那兒。
> Standard: 我記得。 > > Sichuan: 我記倒起的。
G6. 強化 — 「巴適得很 / 安逸慘了 / 扎實」[Hard, Light+]
四川話 favors postposed intensifiers, frequently stacking 「得很」, 「慘了」, 「扎實」 onto adjectives.
| 強化結構 | 範例 | |---|---| | X 得很 | 巴適得很、好吃得很、惱火得很 | | X 慘了 | 累慘了、餓慘了、冷慘了 | | 扎實 X | 扎實好吃、扎實惱火 | | 好 X / 相當 X | 好巴適、相當安逸 |
> Standard: 非常舒服。 > > Sichuan: 巴適得很。 / 安逸慘了。
> Standard: 太累了。 > > Sichuan: 累慘了嘛。
G7. 「老子 / 龜兒」自稱與罵稱 [Hard, Medium+]
「老子」(self-reference, gruff first person) and 「龜兒」/「龜兒子」(2nd or 3rd person, mock-curse) are core 江湖 markers. They are blunt but not necessarily hostile — among friends they are warm; in confrontation they are sharp. Use sparingly in Light, freely in Medium/Heavy.
> Self-reference (gruff): 老子今天累慘了。 > > Mock-curse (friendly): 龜兒子又遲到。 > > Real curse (sharp): 龜兒,你給老子聽倒。
Note: not 1:1 with English curse equivalents. 「老子」 and 「龜兒」 are emotionally calibrated by context and tone, not by intensity alone.
Vocabulary Patterns(詞彙模式)
V1. 標誌詞 — Core Signature Set [Hard, Light+]
The most iconic 川渝 vocabulary. Sprinkle freely — these are not exotica, they are everyday words.
| 川渝詞 | 意思 | |---|---| | 巴適 | 舒服、好、滿意 | | 安逸 | 舒適、爽快、過癮 | | 莫得 | 沒有 | | 搞快點 | 趕緊、快一點 | | 撇脫 | 乾脆、簡單俐落、爽快 | | 擺龍門陣 | 聊天、講故事 | | 牙尖 | 嘴尖、愛說、嘴利 | | 假打 | 裝模作樣、虛偽 | | 鬼火冒 | 火大、氣炸 | | 哦豁 | 糟了、完蛋(驚嘆) | | 板眼兒 | 名堂、花樣、本事 | | 惱火 | 麻煩、棘手 | | 造孽 | 可憐、慘 | | 緊到 | 一直、不停地 | | 慫 | 軟弱、孬 |
V2. 人稱與罵稱 [Hard, Medium+]
| 詞 | 意思 | |---|---| | 老子 | 我(粗豪) | | 龜兒 / 龜兒子 | 罵語 / 戲稱 | | 瓜娃子 | 笨蛋(多戲稱) | | 瓜兮兮 | 傻乎乎 | | 莽夫 | 魯莽的人 | | 傻兒 | 傻子 | | 妹兒 | 女孩 | | 兒娃兒 | 男孩兒 | | 婆娘 | 妻子 / 女人(口語) | | 老漢兒 | 父親 | | 婆婆 | 祖母 |
V3. 文化詞與生活詞 [Hard, Light+]
四川話's lexicon is densely tied to local culture. Use these for setting without leaning on stereotype.
| 詞 | 意思 | |---|---| | 火鍋 | hot pot(共通詞,但語境上極具川渝代表性) | | 麻辣燙 | 麻辣燙 | | 串串兒 | 竹籤穿料的小火鍋 | | 茶館 | 茶館(成都重要社交空間) | | 擺龍門陣 | 聊天 | | 川劇 | 川劇 | | 變臉 | 變臉(川劇技藝) | | 蓋碗 | 蓋碗茶 | | 雞公車 | 獨輪推車(老式) | | 黃桷樹 | 黃桷樹(川渝常見行道樹) | | 巷子 | 巷子 | | 壩壩 | 平地、廣場 | | 院壩 | 院子 |
避免:把「打麻將、火鍋、懶散」做為角色性格的核心;那是刻板印象,不是 dialect。
V4. 程度副詞與口頭禪 [Hard, Light+]
| 詞/套式 | 用法 | |---|---| | 好 X | 好巴適、好辛苦 | | 相當 X | 相當安逸(鄭重強調) | | 巴適得很 | 「非常舒服」 | | 扎實 X | 扎實好吃 | | 莫得辦法 | 沒辦法 | | 哦豁不得了 | 糟糕了 | | 搞快點 | 趕緊 | | 緊到 X | 一直在 X | | 將就 | 湊合 | | 嘿(嗨) | 強化 |
V5. 動詞詞替換 [Hard, Light+]
Routinely swap Putonghua verbs for 川渝 equivalents.
| 普通話 | 川渝 | |---|---| | 看 | 瞅、瞄 | | 走 | 走、跑路 | | 坐 | 杵、坐到 | | 站 | 站起 / 杵到 | | 拿 | 拿起、揣 | | 給 | 拿給、捎 | | 害怕 | 怕、慫 | | 罵 | 凶、剋 | | 吃 | 吃、整 | | 做 | 搞、整、弄 | | 想 | 想、捉摸 | | 不知道 | 莫曉得、不曉得 |
Particle Patterns(語尾助詞)
四川話 particles ride on top of sentences in a denser stack than Putonghua, each with a distinct pragmatic function. Combination rule: usually one particle per sentence; 「嘛」 may chain after 「噻」 / 「撒」 for emphasis.
P1. 嘛 — 緩和 / 理所當然 / 軟性催促 [Hard, Light+]
The most ubiquitous 川渝 particle. Softens commands, marks "that's obvious", or gentle complaint.
> Reasoning-evident: 就是這個道理嘛。 > > Soft urging: 你來嘛。 > > Mild complaint: 莫亂說嘛!
P2. 噻 / 撒 — 肯定 / 祈使 [Hard, Light+]
Marks confirmation or imperative — closer to Putonghua "啊" but with川渝 edge. 撒 is a common written variant.
> Confirmation: 就是這樣噻。 > > Imperative: 過來噻! > > Variant: 拿給我撒。
P3. 嘞 — 完成 / 變化 [Hard, Light+]
Marks change of state or completion, with 川渝 cadence. Roughly = 了 + 啦.
> Completion: 我吃老嘞。 > > State-change: 天黑嘞。 > > Decision: 算嘞算嘞。
P4. 哈 — 確認 / 友善提醒 [Hard, Light+]
Seeks light agreement or marks friendly reminder. Slightly warmer than 噻.
> Agreement-seek: 巴適得很哈? > > Reminder: 路上小心哈。 > > Confirmation: 就這樣哈。
P5. 喲 / 哦 — 驚訝 / 恍然 [Soft, Medium+]
Marks discovery, surprise, mild alarm.
> Discovery: 還真有兩下子喲。 > > Alarm: 哦豁,遭嘞!
P6. 嗦 / 唦 — 詢問 / 推斷 [Soft, Medium+]
Roughly = "is that so?" / asking to confirm an inference.
> Inferential ask: 你不去嗦? > > Confirming inference: 原來如此嗦。
Structure Patterns(結構模式)
S1. 短句 + 反問 [Hard, Light+]
四川話 rhythm is short, fast, blunt. Long subordinated periods are not 四川話 voice; break them down. Rhetorical 反問 is the daily structure of confrontation or surprise.
| 反問套式 | 場合 | |---|---| | 你說啥子? | 沒聽清 / 不爽 | | 咋的嘛? | 挑釁 / 不服 | | 哪個說的? | 質疑來源 | | 你龜兒搞啥子? | 強烈不滿 | | 我搞錯了嗦? | 自我懷疑 |
> Standard: 我並不認為這件事是這樣的。 > > Sichuan: 哪個說是這樣?我看莫得這回事。
S2. 自嘲與江湖氣 [Hard, Medium+]
The 川渝 voice is willing to laugh at itself. Self-deprecation through 「老子」 over-claims ("老子今天又把車鑰匙弄丟嘞") balanced by mock-bragging ("老子這手藝,巴適得很") is the texture of川劇丑角 and《讓子彈飛》張麻子 alike. Use it to de-fuse heavy emotions.
> Tense: 我又把事情搞砸了。 > > 川渝 self-mock: 老子又把這事兒搞砸嘞,瓜娃子一個。
S3. 擬聲與動感 [Soft, Medium+]
四川話 carries strong onomatopoeia and motion-words.
| 擬聲 / 動感 | 場合 | |---|---| | 撲哧 | 笑出聲 | | 啪嗒 | 物件落地 | | 嘩啦 | 水聲 / 倒塌 | | 咕嚕咕嚕 | 火鍋翻滾 | | 唏哩呼嚕 | 吃麵聲 | | 咔 | 突然停 | | 咚咚咚 | 腳步聲 / 心跳 |
S4. 寒暄與起頭 [Soft, Medium+]
| 開場 | 場合 | |---|---| | 哎喲 | 驚嘆、不滿 | | 哦豁 | 糟了、突發 | | 嘿 | 招呼 | | 你龜兒 | 戲謔招呼 | | 走嘛 | 邀約 | | 來來來 | 招呼入座 |
Meta Patterns(後設規則)
M1. 全形標點零違規 [Hard, Light+]
CJK 文本中所有標點一律全形。ASCII 半形 , ; : ? ! 在 CJK 字之間出現一律是 bug,必須換成全形 ,;:?!。對照表:
| 半形(禁) | 全形(用) | Unicode | |---|---|---| | , | , | U+FF0C | | ; | ; | U+FF1B | | : | : | U+FF1A | | ? | ? | U+FF1F | | ! | ! | U+FF01 | | ( ) | ( ) | U+FF08 / U+FF09 |
破折號用兩個全形「—」連寫成「——」;引號用「」/『』;句號用「。」;全形標點與 CJK 之間零空格。
寫完後必跑自檢指令(指令名留作機械驗證,不可省略):
python3 -c "
import re, sys
text = open(sys.argv[1]).read()
bad = re.findall(r'[^\x00-\x7f][,;:?!][^\x00-\x7f]', text)
print(f'half-width punct between CJK: {len(bad)}')
for b in bad[:5]: print(repr(b))
" 輸出檔.md
結果必須為 0。非 0 表示輸出時混入了 ASCII 半形於 CJK 字之間,逐字替換後重跑,直到歸零才算交付。
Step 3: Produce Sichuan Mandarin Output
Apply the checklist to transform the entire text. Aim for a natural 川渝 voice at the chosen intensity. Always check the soul clause first — if the output reads as Putonghua with a few swaps, push 莫/莫得, particles, and 江湖 cadence harder.
Intensity Levels
Each rule is tagged with its minimum intensity (e.g., [Hard, Light+] = active from Light up; [Soft, Heavy] = mainly Heavy).
| Level | Active Rules | Effect | Marker Density | |---|---|---|---| | Light | G1-G4, G6, V1, V3, V5, P1-P4, S1, S4. No 老子/龜兒, no 川劇丑角 cadence. | 普通話 backbone with occasional 川渝 flavor — 巴適/莫得/嘛/哈/搞快點 sprinkled. Like a Sichuanese speaker writing a slightly dressed-up WeChat post. | 5-6 sentences carry 1-2 markers | | Medium (default) | Light + G5, G7 (selective), V2, V4, P5-P6, S2-S3. 老子/龜兒 used in self-reference and friendly banter. | Clearly 川渝 — signature words, particles, blunt cadence, mild self-mockery. Like a 重慶 茶館 conversation transcribed. | every 2-3 sentences carry ≥1 marker | | Heavy | All rules including G7 free, S2 full 江湖 cadence, dense particles, 川劇丑角 self-mockery. | Heavy 川渝 — 張麻子-grade 江湖 voice, every line salted, rhetorical 反問 frequent, 老子/龜兒 freely used. | every 1-2 sentences carry ≥1 marker, ≥2 in dialogue |
If the user doesn't specify, use Medium.
Input-Specific Handling
Mandarin input → Sichuan Mandarin:
- Swap 沒/沒有 → 莫/莫得 systematically (G1).
- Swap 什麼/怎麼/誰 → 啥子/咋個/哪個 (G4).
- Replace key verbs with 搞/整 family (G3) and V5 swaps.
- Add particles 嘛/噻/嘞/哈/撒 (P1-P4) where pragmatic meaning calls.
- Add 兒 化 in legitimate slots only (G2).
- Inject signature lexicon (V1, V4) — aim for ≥3 markers per paragraph at
Medium+.
- Apply intensifier patterns (G6).
- At Medium+, allow 老子/龜兒 in self-reference / friendly banter (G7);
at Heavy, freely.
- Break long sentences into short blunt clauses; insert rhetorical 反問
where it fits (S1, S2).
- Verify the soul clause — if reading the output back doesn't sound
like 川渝 江湖+幽默 voice, push harder on particles, 莫得 negation, and 自嘲 cadence.
Non-Mandarin input → Sichuan Mandarin:
- Translate to Standard Written Chinese first.
- Apply Sichuan transforms as above.
Step 4: Output
- File output: Save to
_sichuansuffix (e.g.,article.md→
article_sichuan.md). If no file was provided, output directly in conversation.
- Show 3-5 representative sentence pairs comparing Original and Sichuan
Mandarin.
- List applied patterns: Note which major pattern groups (G/V/P/S) were
most active.
- Glossary: List signature 川渝 terms used, with Putonghua equivalents.
Anti-patterns(常見失敗模式)
寫完一稿、自檢之前,先掃一眼這條最容易踩的坑:
- 半形標點混入 — 輸出時把
,;:?!打成 ASCII 半形而非全形,;:?!。
常見原因是 IME 切換失誤或從英文 prompt 複製字串。M1 已明文要求全形; 寫完務必跑 M1 自檢指令,計數歸零才算交付。
Important Notes
- This is a stylistic transformation tool for creative, educational, and
cultural purposes.
- 四川話 is a legitimate Mandarin variety (西南官話 成渝片) covering
Sichuan, Chongqing, and adjacent areas of Yunnan, Guizhou, Hubei, Shaanxi — ~120 million speakers, one of the largest Mandarin sub-dialects. Treat it with linguistic respect.
- Linguistic background. Phonology: 入聲 collapsed into 陽平 (the
innovative trait that distinguishes 西南官話 from northern Mandarin); 兒 化 limited to 4 韻尾 ([ɚ][iɚ][uɚ][yɚ]) vs Beijing's 26; 平翹 distinction largely absent. Lexicon: a substantial layer of words are local (巴適/搞快點/莫得) rather than just pronunciation variants of standard Mandarin.
- 重慶 vs 成都 micro-difference. Both belong to 成渝片 and are mutually
intelligible, but the registers differ slightly: Chongqing leans harder / sharper / more 江湖, Chengdu softer / more leisurely / more 茶館-style. Particles 嘛/噻 are more frequent in Chengdu; 老子/龜兒 land more naturally in Chongqing-flavored speech. The skill does not enforce a split — choose register by content, not by city tag.
- Avoid stereotype. Do NOT lean on "lazy / mahjong-addicted / hotpot-
obsessed" tropes when characterizing speakers or scenarios. The dialect's lexicon includes 火鍋 / 茶館 / 川劇 because those are real cultural anchors, not because Sichuanese identity reduces to leisure. The voice is 江湖+幽默 — direct, self-aware, willing to laugh at fate — not idleness.
- 「老子」 and 「龜兒」 are calibrated by context. Among friends they are
warm or playful; in confrontation they sharpen into curse. Don't dump them into Light register — they signal a step toward 江湖 cadence.
- 「莫得」 ≠ random 沒. 莫 is the negator; 莫得 specifically means 沒有
(don't have). 莫得辦法 = no way out, 莫法 = abbreviated form. Use the right one.
- High-frequency markers also include: 要得 (good / ok / agreed —
the signature Sichuan affirmative), 嘎嘎 (meat, child-talk), 扯把子 (chat / gossip / bullshit). Add these to the active vocabulary in Medium/Heavy outputs.
- 「兒」化 is restricted. Only 4 韻尾 in 四川話 vs. 26 in northern
Mandarin. Add 兒 to place names, small objects, playful adjectives, kinship terms — not to abstract nouns or actions.
- Particles are not random filler. Each carries function:
- 嘛 = soft / obvious
- 噻 / 撒 = confirmation / imperative
- 嘞 = completion / change of state
- 哈 = friendly agreement-seek
- 喲 / 哦 = surprise / discovery
- Formal / classical input produces absurd output. A legal contract or
classical poem rendered into 四川話 will be comical, not useful — warn the user before transforming.
- Simplified vs. traditional. This skill follows the input character set
by default — 繁 in, 繁 out; 簡 in, 簡 out. Mainland 四川話 writing is typically simplified; if the user wants authentic mainland feel, ask.
- When in doubt, think: "Would this come out of a Chengdu 茶館 regular's
…
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: tzengyuxio
- Source: tzengyuxio/skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.