Skip to content

majdghithan/agent-skills / skills/lahja

Arabic dialects (لهجة) writing skill - write or rewrite Arabic in a specific spoken dialect with correct dialect GRAMMAR, not just sample words. Covers negation, future and aspect markers, demonstratives, question words, spelling habits, MSA leaks to remove, and how English tech terms sit inside dialect text. Covers Levantine (Palestinian-anchored), Egyptian, Gulf, Iraqi, Maghrebi (Moroccan, Algerian, Tunisian), Libyan, Sudanese, and Yemeni. Use when the user asks to write "بالعامية", "باللهجة الفلسطينية/الشامية/المصرية/الخليجية/العراقية/المغربية/الليبية/السودانية/اليمنية", "in Levantine/Palestinian/Egyptian/Gulf/Khaleeji/Iraqi/Moroccan/Libyan/Sudanese/Yemeni Arabic or Darija", "make this sound less formal/less translated", "convert فصحى to عامية", or writes to you in a dialect and wants dialect output (social posts, chat/WhatsApp copy, casual product copy, scripts). Built on the fasih method by Maher El Gamil. Works standalone; if the separate fasih skill is also available, invoke it for editorial quality while lahja owns the dialect forms. Do NOT use for formal MSA writing, legal/government/medical text, literal translation, or when no dialect is wanted.

FileKindSize
SKILL.mdskill md12.2 KB
references/egyptian.mdmarkdown11.1 KB
references/gulf.mdmarkdown12.0 KB
references/iraqi.mdmarkdown12.5 KB
references/levantine.mdmarkdown16.0 KB
references/libyan.mdmarkdown13.2 KB
references/maghrebi.mdmarkdown15.6 KB
references/sudanese.mdmarkdown11.5 KB
references/yemeni.mdmarkdown12.8 KB
scripts/check_output.pyscript3.7 KB
scripts/check_skill.pyscript4.1 KB

references/iraqi.md

View on GitHub
# Iraqi Arabic (عراقي) - writing guide

**Coverage:** Baghdadi Muslim Arabic (the **gilit** dialect), the media and online standard, with notes for southern (Basra, Amarah, Nasiriyah) and Mosuli/northern qeltu, which differ in ways that break a text if blended. Every form below is cited to the sources at the end; anything the sources didn't settle is listed in the last section and must not be generated.


**Confidence tags:** unmarked = Alkalesi's Georgetown textbook, Albuarabi's dissertation on Iraqi negation, or Wikipedia citing specialists. `(weak)` = Wiktionary or native-answer threads only. Forms we could not source are left out, and section 10 lists what was deliberately dropped.

## 0. Sub-dialect table

| Feature | Baghdadi gilit (default) | Southern (Basra, Nasiriyah) | Mosuli qeltu |
|---|---|---|---|
| "not" + verb | ما | **ما...ش** and مش | ما |
| In progress | دا- | گا- | كِ- / قِ- / عِ- |
| ق | said g, written گ | g | stays q (اقول) |
| ك before front vowels | said ch, written چ | چ | stays ك |
| ر | r | r | merges toward غ (راح -> ʁaħ) |

Blending these is the fastest way to sound fake. Pick Baghdadi unless the user says otherwise. [Albuarabi] [NorthMeso] [SouthMeso]

## 1. Negation

- **ما + verb** (never with an imperative): ما يعرف، ما يريد، ما رحت. [Alkalesi]
- **مو + nouns, adjectives, adverbs** (nominal predicates only): علي مو طالب، مو مشكلة. [Alkalesi] [Wikt-mu]
- **أكو / ماكو** = "there is" / "there isn't": ماكو أكل بالثلاجة. **ماكو** also works as a prohibition before a verb: ماكو تخبرني كل نص ساعة. [Wikt-aku] [Albuarabi]
- **Possession**: ما عندي، ما معك. [Alkalesi]
- **Prohibition**: لا + imperfect: لا تسافر، لا تسافرين، لا تسافرون. [Alkalesi]
- **ولا** links further negated items: ما أجي ولا أروح. [Alkalesi]
- **Negation sits outside the aspect marker**: ما دا يدرس, not ~~دا ما يدرس~~. [Albuarabi]
- **With a demonstrative**: مو goes **before** it when the noun is definite (مو هالبيت), **after** it when indefinite (هذاك مو بيت). [Alkalesi]
- **No ...ش in Baghdadi**, but **southern Iraqi does use ما...ش and مش**. Don't write ما أعرفش in a Baghdadi text, and don't tell a Basrawi it isn't Iraqi. [Albuarabi]

## 2. Future

- **رح- or حـ- + imperfect**: رح أدرس، رح يقرا كتاب، حـ- often feels more immediate. راح is a variant; Alkalesi recommends the short رح. [Alkalesi]
- Before a two-consonant start, a helping vowel appears: رح إنسافر. [Alkalesi]
- **راح also works as a standalone auxiliary**: أحمد ما راح يروح للجامعة. [Albuarabi] [Wikt-rah]
- **Negated future**: ما راح + verb. [Albuarabi]
- **No بـ future.** Iraqi بـ / بي- is the preposition "in, with", so a Gulf-style بروح collides with it. [Alkalesi] [Persson]

## 3. Aspect - the single most important Iraqi rule

- **Bare imperfect = habitual**, not "right now": نشرب چاي = we drink tea. [Alkalesi]
- **دا- + imperfect = happening now**: دا نشرب چاي = we are drinking tea. More pairs: آكل خبز vs دا آكل خبز; وين تروح كل يوم؟ vs وين دا تروح اليوم؟ [Alkalesi]
- **An agent that writes the bare verb for "right now" sounds wrong in Iraqi.** This is the highest-value rule in this guide.
- **But دا- marks an action already under way, not a future one.** With رح / حـ, or after a modal, use the bare imperfect even when the sentence says "now": رح أشوفه هسّه, **not** ~~دا أشوفه هسّه~~. هسّه in a future sentence does not license دا-.
- **گاعد + imperfect** is a fuller progressive (more southern/Najafi in the sourced data); دا- is the Baghdadi default. [Albuarabi]
- **Regional forms of the same marker**: Baghdadi دا-، Najafi/southern گا-، Mosuli كِ-/قِ-/عِ-. [Albuarabi]
- **No بـ indicative prefix** anywhere in Iraqi. Don't import Levantine بكتب. [Alkalesi] [Persson]
- **Bare imperfect after modals**: أريد أروح، لازم أشوف، أگدر أصرّف. Note أريد is genuinely Iraqi here, not a formal-Arabic leak. [Alkalesi]

## 4. Demonstratives and question words

- **Near**: هاذا m، هاذي f (colloquial written هاي)، هذولا pl. **Far**: هذاك، هذيچ، هذولاك. [Alkalesi] [Wikt-hay]
- **هـ- before a definite noun is the more common usage**, assimilating to sun letters: هالولد، هالبيت، هالسيارة، هالطالبة، and with an English noun هالـ query. [Alkalesi]
- **Question words**: شنو (also the proclitic شـ-: شسمك؟ شدا تسوّي؟)، وين، منين/إمنين، ليش (variants الويش، لويش)، شلون، شوكت، منو، شگد، بيش، چم/كم، أي / يا / ياهو. [Alkalesi] [Wikt-shnu]
- **شكو ماكو؟** is the standard friendly greeting ("what's up"), answered with ماكو شي اليوم. [Alkalesi] [Wikt-shaku]
- **هل** exists but Alkalesi marks it as barely used. Ask by intonation. [Alkalesi]
- **Relative اللي** is invariable and only used with a **definite** head: الأستاذ اللي يدرّس عربي. With an indefinite head it simply disappears: أكو رجّال يريد يسلّم عليك. [Alkalesi]

## 5. Pronouns

آني (I, no gender split)، إنتَ، إنتي، هو، هي، إحنا، إنتو، همّه. There is **no dual and no separate feminine plural**: إنتو and همّه cover both, with gender showing on the adjective (إنتو زينين / إنتو زينات). **No copula**: آني زين = "I am well". [Alkalesi]

Suffixes: ـي / ـني، ـنا، ـك (2ms)، **ـچ (2fs: كتابچ، شلونچ)**، ـكم، ـه، ـها، ـهم. The ـچ is a gilit signature; Mosuli qeltu keeps ـك. [Alkalesi] [Wikt-shlon]

## 6. Spelling habits

- **ق said as g is written گ** in dialect writing: گال، گلت، أگدر، شگد. Keep ق for MSA-layer and loan words. This differs from Levantine and Egyptian, where ق stays ق. [Wikt-gal] [BaghdadWiki]
- **ك said as ch is written چ**: چان، چم، چذي. [Wikt-chan] [Alkalesi]
- **Interdentals ث / ذ / ظ are pronounced and kept.** Do not "colloquialise" them into ت / د / ض the way urban Egyptian and Levantine writing does. [BaghdadJewishWiki] [Albuarabi]
- **دا- is usually written as a separate دا** before the verb: دا آكل خبز، دا أحچي. A joined داكتب also circulates; pick one. (weak, native-answer thread) [HiNative]
- **"now"**: هسّة is the etymological spelling (from هالساعة), **هسّه is what people actually type**. [Wikt-hassa]
- **پ and ڤ** appear only in foreign words and often collapse to ب / ف. [BaghdadWiki]
- **No tanween, no short-vowel marks.** There is no published Iraqi CODA, so dialect spelling online is inconsistent by nature: pick your conventions and hold them across the text. [TunCODA]

## 7. MSA leaks - kill on sight

| Leak (MSA) | Iraqi | Source |
|---|---|---|
| سوف / سـ | رح- / حـ- | [Alkalesi] |
| لن | ما راح | [Albuarabi] |
| لم | ما + past | [Alkalesi] |
| ليس | مو | [Alkalesi] [Wikt-mu] |
| الذي / التي / الذين | اللي (definite heads only) | [Alkalesi] |
| هذه | هاذي / هاي | [Alkalesi] [Wikt-hay] |
| ماذا | شنو, or the proclitic شـ- | [Alkalesi] |
| لماذا / أين / كيف / متى | ليش / وين / شلون / شوكت | [Alkalesi] |
| الآن | هسّة (informal هسّه) | [Wikt-hassa] |
| جداً | كلش (intensifies adjectives) | [Wikt-kullish] |
| كثير (quantity) | هواية (nouns and adverbs) | [Wikt-hwaya] |
| فقط | بس | [Wikt-bass] |
| لا يوجد | ماكو | [Albuarabi] |
| يستطيع | أگدر / يگدر | [Alkalesi] |
| يفعل | يسوّي | [Alkalesi] |
| قام بـ + verbal noun | plain verb | style rule, see SKILL.md |

**كلش and هواية are not interchangeable**: كلش = "very" on an adjective, هواية = "a lot / many" on a noun. They can stack: أحبچ كلش هواية. [Wikt-kullish] [Wikt-hwaya]

## 8. English tech terms

Iraqi Arabic-English code-switching is documented (WhatsApp and student-conversation studies; a paper on computer, internet, and mobile-phone loanwords), and single English nouns inserted into an Arabic sentence are the most common pattern. [Turki] [Salman17]

- **Keep English nouns in Latin script** and attach Arabic clitics before them: الـ query، هالـ build.
- **English loan nouns take Arabic plurals** in attested Iraqi data (the sound feminine plural on borrowings). [Saeed10]
- **Not verified for Iraqi:** English verbs taking Iraqi verb morphology, and any سوّى + English noun pattern. Until it is sourced, write an Iraqi verb plus the English noun as its object and keep it simple.

## 9. Worked rewrites (MSA -> Baghdadi)

Grammar forms come from the sections above; ordinary content words are not individually sourced.

**1. Progressive vs habitual** (الآن, ماذا)
- MSA: ماذا تفعل الآن؟
- Iraqi: شنو دا تسوّي هسّه؟
- The bare شنو تسوّي؟ would mean "what do you do (in general)".

**2. Future + negation** (سوف, لن, يستطيع)
- MSA: لن أستطيع أن أحضر الاجتماع، وسوف أكتب التحديث بعده.
- Iraqi: ما رح أگدر أجي للاجتماع، ورح أكتب الـ update بعدين.

**3. Existential + nominal negation** (لا يوجد, ليس)
- MSA: لا يوجد أي خطأ في الـ log، لكن الـ build ليس جاهزاً.
- Iraqi: ماكو أي error بالـ log، بس الـ build مو جاهز.

**4. Question + intensifier** (لماذا, جداً, demonstrative prefix)
- MSA: لماذا هذا الـ query بطيء جداً؟
- Iraqi: ليش هالـ query كلش بطيء؟

## 10. Deliberately NOT included (unverified)

Do not generate these, they are not attested in the sources checked: the negative-copula paradigm ماني / مانت / ماهو (the table circulating for this is **Moroccan**, and Iraqi sources say the southern group has no negation pronouns); مب / موب (Gulf, no Iraqi section in the sources); ودّي as a future marker (it is volitional, "I wish", and unsourced for Baghdadi); شمعنى for ماذا; يمتى for متى; هم / بعد for أيضاً; a دا- imperative (دا روح); a distinct ـهن feminine plural suffix. [Albuarabi] [Alkalesi]

## Feedback

**Found something wrong or missing?** Arabic dialects vary by city, generation, and family, and this guide will not match every speaker. Tell Majd: majd.ghithan20@gmail.com, or open an issue or PR at https://github.com/majdghithan/agent-skills.

## Sources

- [Alkalesi] Alkalesi, Modern Iraqi Arabic: A Textbook (Georgetown UP) - https://archive.org/details/modern-iraqi-arabic-a-textbook-by-yasin-m-alkalesi
- [Albuarabi] Albuarabi 2021, The Syntax of Negation in Iraqi Arabic (UW-Milwaukee) - https://minds.wisc.edu/bitstreams/fb8708b2-725f-4b2f-8897-d9ff3441f37c/download
- [MesoWiki] https://en.wikipedia.org/wiki/Mesopotamian_Arabic · [BaghdadWiki] https://en.wikipedia.org/wiki/Baghdadi_Arabic
- [NorthMeso] https://en.wikipedia.org/wiki/North_Mesopotamian_Arabic · [SouthMeso] https://en.wikipedia.org/wiki/South_Mesopotamian_Arabic
- [BaghdadJewishWiki] https://en.wikipedia.org/wiki/Baghdad_Jewish_Arabic
- [TunCODA] Dialectal Arabic orthography inconsistency - https://arxiv.org/pdf/2402.12940
- [Persson] b-prefix scope (Gulf and Levantine, not Iraqi) - https://www.lancaster.ac.uk/jais/volume/docs/vol8/persson_web.htm
- [Salman17] Salman & Mansoor 2017, English loanwords in Iraqi Arabic - https://cuesj.cihanuniversity.edu.iq/2017/06/01/vol1no1_e5/
- [Turki] Turki & Sabti, WhatsApp code-switching among Iraqi speakers - https://www.academia.edu/40835902/
- [Saeed10] Sa'eed 2010, Mosuli Arabic loan plurals (cited in the above)
- [HiNative] Native answers on the دا prefix and its spelling - https://hinative.com/questions/18917778
- Wiktionary: [Wikt-aku] https://en.wiktionary.org/wiki/أكو · [Wikt-mu] https://en.wiktionary.org/wiki/مو · [Wikt-rah] https://en.wiktionary.org/wiki/راح · [Wikt-hay] https://en.wiktionary.org/wiki/هاي · [Wikt-shnu] https://en.wiktionary.org/wiki/شنو · [Wikt-shaku] https://en.wiktionary.org/wiki/شكو · [Wikt-shlon] https://en.wiktionary.org/wiki/شلون · [Wikt-gal] https://en.wiktionary.org/wiki/گال · [Wikt-chan] https://en.wiktionary.org/wiki/چان · [Wikt-hassa] https://en.wiktionary.org/wiki/هسة · [Wikt-kullish] https://en.wiktionary.org/wiki/كلش · [Wikt-hwaya] https://en.wiktionary.org/wiki/هواية · [Wikt-bass] https://en.wiktionary.org/wiki/بس
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall