Writing & Editing

Transforma notícias, comunicados de bancos, Procons ou polícias e descrições de golpes digitais (falsa central, falso parente, voz clonada por IA, falso INSS, boleto falso, Pix, QR code, anúncio de empréstimo) em alertas curtos em linguagem simples para pessoas com 60 anos ou mais e para quem cuida delas, com um golpe, um sinal, uma ação e uma frase para lembrar, mais uma versão para WhatsApp. Também revisa alertas já escritos contra as técnicas do art. 5º da Lei 15.263/2025 (Política Nacional de Linguagem Simples). Use sempre que alguém pedir para explicar, resumir, traduzir ou divulgar um golpe para idosos, pais, avós ou o grupo da família, mesmo sem dizer "linguagem simples".

Professional localization agent for translating macOS applications, developer tools, audio engineering software, and system-level UI across Tamil (ta_IN), English (en_US, en_GB, en_AU), Spanish (es_ES, es_419), and German (de_DE, de_AT, de_CH). USE THIS SKILL whenever the task involves: translating UI strings, localizing .strings/.stringsdict/.xliff files, reviewing translations for Apple platform apps, writing error messages or alerts in multiple languages, handling audio/system terminology in non-English locales, checking Tamil purity (no English loanwords in Tamil script), German compound noun correctness, Spanish regional sensitivity (Spain vs LATAM), or English variant spelling (US/UK/AU). Also trigger when the user mentions localization, i18n, l10n, translation for macOS/iOS, Apple HIG compliance in other languages, or needs culturally correct technical translations. Even partial overlap with multilingual UI string work should trigger this skill.

Build a 14-to-30 day Hebrew marketing plan and content calendar (SWOT, USP, influencer strategy, and a full day-by-day calendar with hooks and copy) for a agency client, delivered as a Word (.docx) + PDF twin. Use whenever a agency employee asks to build a "תוכנית שיווק", "תוכנית תוכן", "לוח תוכן" / content calendar, or says "תבנה לי תוכנית שיווק ללקוח", "תכין לוח תוכן לחודש", "content calendar for [client]", or asks for a posting schedule with real copy and hooks ready to publish. Final step of the agency pipeline — always requires a client profile (client-profiler) AND competitor research (competitor-research) to already exist; if either is missing, stop and offer to produce it first instead of generating a generic plan. Not for one-off single posts or single-campaign ad copy — that's simpler in-chat work.

中文學術潤稿站。Use when 中文潤稿、去中文 AI 味、chinese-academic-editing。

Reviews technical docs against the code they describe and reports where they disagree, and whether a developer or an agent could act on them. Returns findings only and never fixes anything. Use whenever the user asks to review, audit, check or verify a README, docs site, docs folder, API or CLI reference, guide, changelog or agent instruction file, asks whether docs are accurate, up to date or match the code, or asks what has drifted. The impa skill also starts this automatically after every round. Do not use for marketing pages (use beedle-review), for artifact pages or a claude.ai link (use mudora), for app UI (use ultrahand:review) or for code review.

Writes and updates the technical docs that live with a project's code, checks every command, flag, path and default against that code, then starts an independent review without being asked. Use whenever the user wants a README, docs site or docs folder, API or CLI reference, quickstart, tutorial, how-to guide, troubleshooting page, changelog, migration guide, runbook, decision record, contributing guide, or an instruction file for agents such as AGENTS.md or CLAUDE.md, or asks to update, restructure or fix any of these, even if they only say "document this" or "the docs are out of date". Repository docs belong here even when the user says "doc". Do not use for marketing or landing pages (use beedle), for artifact pages, reports, plans, handoffs or decks (use mudora), or for reviewing docs without changing them (use impa-review).

英文學術稿件的潤稿站：清除 AI 生成痕跡（AI tells）、對齊學術慣例、貼合作者與期刊語氣， 一律不動數字、標記、citekey、claim strength、validation boundary 與方法／限制語意。 Use when 英文潤稿、潤飾英文稿、去英文 AI 味、English academic editing、de-AI the English manuscript、polish the English draft，或在 academic-pipeline 的 translate-en 之後、英文 人工確認之前。內容源自 academic-humanizer v0.3.3 的 Layer 1/2/3/5；claim-evidence 改寫 （原 Layer 4）刻意不納入，改由上游 claim-audit 負責；提案模式（原 Layer 6）另出獨立 skill。

Clean AI-generated drafts in English or Spanish. Removes chat residue and machine artefacts, flags or removes invented or unsupported content, fixes spelling and style-guide errors, and tones down AI habits only where they repeat or say nothing, while keeping the writer's language variety, tone, genre and register. Use for "humanize this", "humanizar", "que no suene a IA", "limpiar este borrador" or "quitar el tono de IA".

Establishes a project's quality bar as a written contract and stops agents quietly lowering it. Interviews the user on which dimensions matter, supplies sane default thresholds when they have no number in mind, records everything in CONSTRAINTS.md, and watches the diff for a weakened bar — new @ts-ignore or eslint-disable suppressions, skipped or deleted tests, assertions stripped out, unimplemented stubs, thresholds edited down. Use when no quality bar is written down, when the user says "set up constraints" or "define our standards", when the user wants dimensions they care about — accessibility, web performance, coverage — set up as enforced constraints, when an agent keeps silencing checks or skipping tests to get to green, when you need a coverage or performance threshold and don't know what number to pick, or when an agent writes more code than anyone will read.

Evidence-first writing for technical documentation: READMEs, architecture and design docs, ADRs, API contracts, security and compliance claims, release notes, audit packs, and handoff notes. Use whenever a document states that something is secure, correct, complete, verified, tested, certified, compliant, or guaranteed; whenever you review documentation for unsupported claims; and whenever prose must survive a reviewer, an auditor, or an incident. Every material claim gets a mechanism, evidence, scope, and status.

Builds a brand voice and tone guide and a messaging house from your Google Ads, Meta ads, social posts and site copy, then checks live copy against it. Use when you need a style guide, a messaging framework or a website messaging audit. Not for writing new copy (use copywriting)

Provider-independent character design and continuity direction for generated media. Use when an agent must create, preserve, revise, or QA a recurring character across AI-generated images, video clips, avatars, animation, comics, ads, product mascots, serialized social posts, or multi-shot campaigns; especially when building character bibles, reference packs, prompts, expression/pose ranges, costume/hair/makeup/prop continuity, identity and likeness safeguards, iteration plans, handoff notes, or continuity reviews.

Use when the villain is more compelling than the hero, a villain feels flat or cartoonish, you are writing an antihero, or you want a villain's fall to feel earned. Gives patterns for where villains and antiheroes sit on the three scales and how to build them.

Photoreal character asset sheet (三视图) prompt generator for Nano Banana Pro — one text-to-image prompt per character, no reference image: right panel head-and-shoulders close-up, left panel invisible-mannequin front full-body + back full-body. Uses a fixed verbatim template with ultra-detailed face description, a single identifying mark, a cinematic-portrait hyperreal style block plus per-character skin / hair / fabric lines, and no shallow depth of field. Use whenever the user wants a character asset, character sheet, character turnaround, model sheet, 三视图, 人物三视图, 人物资产, 角色资产, 人物设定图, 角色设定图, 定妆照, or says '帮我做个人物' / '出一个角色' / '给这个人设写提示词' — for any project. 触发：人物资产、角色三视图、人物设定图、character sheet。

Use when deciding what a character is good at, when a hero is either helpless or too perfect, when a competent character reads as dull or arrogant, or when strength is not about fighting. Gives rules for designing skills, strengths, and weaknesses that drive the plot.

Draft, restructure and polish experimental or phenomenological high-energy-physics papers (hep-ph, hep-ex, ML for HEP) in the clear, story-driven style of five papers by Daniel Whiteson's group, with one central question, ideas layered from familiar to new, plain sentences with the verb up front, results as numbers against a baseline, and figures and captions that carry the evidence. Use it whenever the user drafts or revises any part of a physics paper, thesis chapter or proceedings (title, abstract, introduction, methods, results, conclusions, captions, figure plans, outlines), asks to make physics writing clearer or better organized, says a draft has no story, wants notes (including notes written in Chinese) turned into English paper text, or asks for feedback on a draft or a figure, even without naming this skill.

Turns board or committee meeting notes or transcripts into draft minutes and an action tracker. Records motions, votes, decisions and assignments in standard minutes form, flags anything unclear instead of guessing, and drafts follow-up reminders for review. Use for board meetings, committee meetings, minutes, action items or meeting follow-up, even for one piece. Drafts only; the secretary approves. Not legal advice.

Use whenever the user asks to write, create, update, refresh, rewrite, or audit a README — including plain requests like 'write a README for this repository'. Builds the root README.md from verified code, manifests, scripts, tests and workflows, and never invents commands, badges, configuration, or claims.

Use when writing, editing, or reviewing prose in a Dogmatiq project, in markdown documentation such as READMEs, ADRs, glossaries and changelogs, or in doc comments in code in any language. Triggers: style review, check style, prose style, markdown style, check docs, documentation style, style guide, doc comments, spelling, US English.

Runs a rigorous, evidence-based audit of a target — a script, a doc, a config, a whole project, an enterprise cloud setup, anything — by constructing concrete failure-scenario "thought experiments" and verifying every claim the target makes about a dependency's behavior against a live/authoritative source rather than memory, including claims one paired doc (a README) makes about another (its SKILL.md) and formatting-convention consistency where one is established. Asks once how fixes should be authorized among four CRUD options (blanket, process, file, or every CRUD step — Read is always free, only Create/Update/Delete get gated) before applying any, then re-runs the same audit against the fixed state to catch regressions and closes with one consolidated report. Triggered by the phrase "thought experiment." Iteration count is open-ended.

Rules and a test method for writing titles, meta descriptions, excerpts and H1s that AI agents pick and cite when answering a user's question. Answer-shaped headlines beat clever, problem-first, credential-first and year-stamped ones, and agents back one page per topic per site. Use whenever writing or refreshing a title or meta description, auditing a site's titles for AI search, or picking which of several overlapping posts should win a topic. Trigger on "title and meta", "rewrite the headline", "agent pass", "why isn't this post cited", "which post should own this topic".

Draft, rewrite, or review reader-facing prose so it is specific, useful, and recognizably the author's without inventing facts or performing humanness. Use whenever writing or revising prose that people will read under the author's name, such as social and LinkedIn posts, READMEs, announcements, launch copy, essays, articles, newsletters, documentation, talks and important emails, and for any prose that feels generic, hollow, or AI-shaped or shows AI tells such as not-X-but-Y contrasts, one-line closers, forced triads, dashes everywhere, stock AI words, or bold labels. Supports co-write, rewrite, review, and lint modes.

Academic writing style toolkit, bilingual CN/EN: paper rewriting with AI-flavor self-check and style naturalization (scan reports are detection only), deterministic cleanup (model artifacts, punctuation, filler phrases), integrity guardrails (numbers, DOIs, PMIDs, terminology must survive untouched), and agent-guided revision for clearer, more natural academic prose. Built for scholars whose formal writing reads stiff, templated or machine-flavored — especially non-native (ESL) authors. Skip when: pure grammar polish, full-text translation, or rewriting unrelated to academic writing — use a dedicated tool. Includes AIGC-disclosure compliance checks (China 2025-09 labeling rules) and academic-integrity guardrails: this tool improves writing quality for self-review; it does not help misrepresent authorship, conceal required AI disclosure, or defeat integrity review. 100% local, zero upload, Python stdlib only. Related: paper-polisher-pro (broad polishing), pubmed-verifier, cite-holmes, academic-figures.

Turns messy program data (CSV or Excel) into a funder-ready impact report. Profiles and cleans the data, computes agreed metrics with a script, suppresses small groups, drafts the narrative and checks that every number in the draft matches the computed results. Use for impact reports, funder reports, outcome summaries or program dashboards, even if asked only for a summary. Aggregate data only. Not for client-level case files.