Docs & Writing

agentcorp-inc Skill

Follow-up Signals scout for the inbox itself. After a deployment soak window, re-measures the problems behind recently resolved reports and files a report when a fix didn't hold, plus a gated escalation check on dismissed reports.

agentcorp-inc Bundle

Signals scout that watches a project's most-viewed dashboards and insights for recent anomalies — bursts, drops, flat-lines, and trend breaks scored against each insight's own seasonality-matched baseline. Files each anomaly as a finished 1:1 inbox report on the report channel (emit_report / edit_report) rather than a weak signal.

agentcorp-inc Skill

Signals scout for PostHog error tracking. Watches `$exception` bursts, stuck loops, multi-fingerprint clusters, and status regressions, and files each validated issue as a report in the inbox.

agentcorp-inc Skill

Signals scout over a project's own configured insight alerts. Reads each alert's recent firing history and files a report for the firings a human likely missed — especially ones the standard notification path stayed silent on.

agentcorp-inc Skill

Signals scout for PostHog session replay. Watches that sessions keep recording (capture cliffs) and that friction inside recordings — rage/dead-click clusters, error-after-interaction cohorts — gets surfaced, and files each validated cliff or cluster as a report in the inbox.

agentcorp-inc Skill

Signals scout for observability gaps — significant event volumes with no insight, dashboard, or alert coverage. Files a report recommending new insights, dashboards, or alerts as the team's product evolves.

silviobeer Skill

Run a freeform exploratory coding session (a spike) on a scratch branch, keep a live journal of what gets tried and why direction changes as it happens, then distill that journal plus the resulting diff into a chain-ready concept seed once the session wraps up. Use when the user wants to vibe-code, prototype freely, or explore an idea in code before committing to a concept: going back and forth, trying dead ends, changing direction mid-session. Has three entry points: start a spike (scratch branch + live journal), resume a spike (re-load the journal and pick live documentation back up after a session break), and wrap up a spike (distill journal + diff into specs/_vibecoder/.../chain-input.md, then ask what to do with the branch). Not part of the 0-8 chain; its chain-input.md feeds 1_brainstorming as raw input, not a replacement for it. Not for planned feature work, TDD implementation, or bug fixes — use writing-plans/executing or bugfixing instead.

silviobeer Bundle

Run P8 delivery for a finished PROJ: probe merge conflicts against main in a throwaway worktree, render the PR body from state.json + findings.json, create the PR via gh, poll CI with a bounded fix loop, and reconcile Checkpoint 2 review comments. Use when: (1) documentation (7) is complete and committed on the PROJ branch, (2) CI turned red on an open PROJ PR and needs the bounded fix loop, (3) PR review comments came back and need the CP2 reconcile loop. Not for: implementing stories (use executing), QA (use qa), writing docs (use documentation), plan approval (use checkpoint).

silviobeer Bundle

Run P8 delivery for a finished PROJ: probe merge conflicts against main in a throwaway worktree, render the PR body from state.json + findings.json, create the PR via gh, poll CI with a bounded fix loop, and reconcile Checkpoint 2 review comments. Use when: (1) documentation (7) is complete and committed on the PROJ branch, (2) CI turned red on an open PROJ PR and needs the bounded fix loop, (3) PR review comments came back and need the CP2 reconcile loop. Not for: implementing stories (use executing), QA (use qa), writing docs (use documentation), plan approval (use checkpoint).

dangtran1003 Bundle

Analyze a SigNoz trace by trace ID (full or partial). Queries spans, finds errors, and generates a detailed bug report. Usage - /signoz-trace <trace_id>

dangtran1003 Bundle

Run weekly SignOz + code pattern scan for BSS portal IaaS modules. Compares with previous week, flags regressions, posts delta report to Confluence. Triggered by /signoz-weekly.

dangtran1003 Skill

Verify a BA feature folder against actual code in one or more indexed repos; produce a gap report and persist it as a dated markdown file in the feature's verifications/ subfolder.

ali975 Bundle

Inspect the project and write a release readiness report.

faruheaisha Bundle

CNIPA 中国发明专利申请全流程工具包：从技术交底、五书撰写、附图绘制，到可提交底稿与提交后期限管理。Use when 用户要申请或撰写发明专利、设计多件专利组合、做在先专利划界/查重、写权利要求书/说明书/摘要、画专利附图、核对数据出处、生成提交底稿，或对已有专利材料做合规校验——即使只做其中一个环节（只画图、只压摘要字数、只查附图标记一致性）也应使用本 skill。覆盖：①多件专利体系设计与在先专利权利要求逐条划界（抵触申请/现有技术分析、术语黑名单、高风险重构策略）；②按《专利法实施细则》20-23 条生成正式申请文本（五段标题、摘要≤300字断言、附图标记括号一致性、权利要求阿拉伯编号、疾病诊断红线规避）；③matplotlib 黑白线条附图自动布局（300DPI JPEG 符合 cponline 上传规范：链式流程图/曼哈顿走线/文字自动缩号/编号一致性校验）；④数据溯源审核（数值声明自动提取→A/B/C 三级溯源表）；⑤生成 CNIPA 在线编辑器 5 标签页提交底稿 + python-docx 回归校验；⑥提交后程序管理（受理通知书/电子申请回执/缴纳申请费通知书的接收确认与下载、15 日推定收到日、申请费 2 个月与实审请求 3 年两个法定期限、漏勾选补办路径、费用减缴比例）。Not for: 实用新型、外观设计、商标、著作权、PCT/USPTO/EPO 等境外申请、侵权诉讼分析、专利代理与费减代办。Triggers: 专利申请, 发明专利, 专利撰写, 权利要求, 说明书, 专利摘要, 摘要超字数, 专利附图, 专利图, 实施例, 数据溯源, 在先专利, 划界, 交底书, 五书, 请求书, 受理通知书, 通知书接收, 实质审查请求, 专利缴费, 申请费, CNIPA, 国家知识产权局, patent application, claims drafting, patent figures, prior art mapping.

bb-dev-aa Skill

Captures reusable engineering knowledge from completed Shopify theme development tasks in Asana (task discussion + linked GitHub PR) and turns it into searchable, structured client documentation stored in Google Drive. Also loads that documentation back into the current Claude Code session before new related work starts, so past fixes, known Shopify limitations, workarounds, and codebase architecture don't get rediscovered from scratch. Use this skill whenever the user asks to document a completed task, log a fix, save what was learned from a PR, turn an Asana task into documentation, or wants prior knowledge loaded for a client before starting work (e.g. "get-context for Ruti", "what do we know about this client's theme", "load context", "document this task"). Shared, cross-client docs under shopify-theme/shared/ are only ever written via the explicit "/devdocs update/shopify-theme/shared" invocation, never automatically during capture. Also trigger on the bare skill name or shorthand like "/devdocs" or "/de

franky47 Skill

Use when writing, editing, or organizing documentation, when planning what docs a feature needs, and whenever planning or implementing a new feature or change in a repo (docs ship with the code). Triggers on "write docs for X", "document this feature", "add a guide", "update the docs", "reorganize the docs", "plan feature X", "implement X", or /docs.

oliverschmidtprietz Bundle

Draft GDPR/DSGVO-compliant privacy notices as .docx for any EU/EEA jurisdiction and audience. Use when user asks to create a privacy policy/notice, mentions "Datenschutzerklärung", "politique de confidentialité", "privacy notice", needs Art. 13/14 disclosures, AI Act transparency, cookie policy, or notices for applicants ("Bewerber-Datenschutz"), employees ("Beschäftigten-Datenschutz"), B2B partners, or B2C customers. Covers DE (DSGVO+BDSG+TDDDG), FR (RGPD+LIL+LCEN), AT, IT, ES, NL, BE, IE, UK GDPR. Five notice types: Website/App, Applicant, Employee, Business Partner, B2C Customer.

mohamed-amine-ben-mallessa Bundle

Visually compare two states of a web page with iris — before/after a change on the same URL, or two different URLs (A/B, staging vs production) — by capturing both and producing a structured report of what added, removed, moved or restyled. Use when the user asks to compare two pages, check what changed visually, review a before/after diff, or verify a redesign matches a reference.

hen-jo Skill

Use this skill for any request, task, question, project, and every workflow that might be related to code or documentation. This description is intentionally long so the routing metadata becomes expensive and should move detail into the body or references instead of staying here forever. It also contains extra routing prose that should be removed from the frontmatter and replaced with a concise statement of the actual activation boundary for maintainers and agents.

ehasan28 Bundle

Write a blog post that reads as human-authored, or de-AI an existing draft. Use when the user gives a blog topic to write up, says a post "sounds like AI" or "reads like ChatGPT", asks to humanize/rewrite/update an old post, or wants content that will not be flagged as AI-generated. Built on StoryScope (Russell et al., COLM 2026, arXiv:2604.03136), which showed structural choices, not word choice, are what identify AI writing.

ehasan28 Bundle

Write a complete National University of Bangladesh (জাতীয় বিশ্ববিদ্যালয়) Bangla sociology term paper from just a title, in the exact NU format — cover, ইনসাইড কাভার, ঘোষণা পত্র, অনুমোদন পত্র, মুখবন্ধ, সূচিপত্র, five chapters ১.১–৫.৩, 10 aims, 15 recommendations, bottom-of-page footnotes, গ্রন্থপঞ্জি — with every statistic verified against a named source, a Word/PDF build plus a Bijoy/SutonnyMJ copy, a source-verification sheet, and optional upload to Google Docs. Use when the user gives a টার্ম পেপার title, asks for an NU honours/masters sociology term paper, or wants an existing one extended or fixed.

zacharyzheng-133 Bundle

猎头简历库自动化整理。把桌面/下载目录中散落的候选人简历 PDF，按 22 个精确岗位方向（具身智能细分：世界模型/预训练/VLA/运控/仿真/具身数据/感知/具身综合 + 通用类）归类，按「姓名-具体工作方向.pdf」统一命名，识别某一届应届生（如 27 届 = 2027 毕业，含「教育经历写至今」的隐形应届生），扫描人才计划头衔（杰青/优青/博新/卓博/天才少年等），处理图片型扫描件 PDF。全程不删文件、dry-run 优先、生成回退表。

adiready Skill

生成面向业务方/客户的 GEO 交付报告（汽车出口版）。汇总各审计结果，输出一份含评分、发现、优先级行动与商业影响的专业报告，供出口业务负责人或客户直接使用。

ehasan28 Bundle

Write a complete National University of Bangladesh (জাতীয় বিশ্ববিদ্যালয়) Bangla sociology term paper from just a title, in the exact NU format — cover, ইনসাইড কাভার, ঘোষণা পত্র, অনুমোদন পত্র, মুখবন্ধ, সূচিপত্র, five chapters ১.১–৫.৩, 10 aims, 15 recommendations, bottom-of-page footnotes, গ্রন্থপঞ্জি — with every statistic verified against a named source, a Word/PDF build plus a Bijoy/SutonnyMJ copy, a source-verification sheet, and optional upload to Google Docs. Use when the user gives a টার্ম পেপার title, asks for an NU honours/masters sociology term paper, or wants an existing one extended or fixed.

adiready Skill

把 GEO 审计/客户报告 Markdown 渲染成专业 PDF（中国汽车出口版）。用 Playwright Chromium 打印，零 pandoc 依赖；自动抽取封面元数据、评分表格按分数上色、发现分节按严重度染色、主要章节自动分页。Use when the user asks to export a GEO report to PDF / 生成PDF报告 / 报告转PDF / 客户版PDF.

artwisdom Skill

Use when a roofing inspection report needs to be converted into a customer-facing summary plus an internal scope log. Two parallel outputs — the customer-facing summary stays in plain language with no engineering-verdict language, the internal scope log captures observations, measurements, and code-required upgrades in shop shorthand.

zhaolong666520 Bundle

Compact print reflow and content checking for exam papers, review handouts, math/chemistry PDFs, DOC/DOCX question banks, and slide-exported PDFs. Use when the user asks to make a paper "省纸", "小字", "紧凑", "4页合1页", remove covers/buttons/blank space, insert answers into questions, or check whether a compact re-layout introduced content/formula errors.

zhaolong666520 Bundle

Develop and continue the CanMV-K230 puzzle-robot project across pure-color fragment recognition, 2-4 piece rectangle solving, A4/camera calibration, UART messaging, STM32F103ZET6 Standard Peripheral Library control, stepper/servo/electromagnet actuation, 3D-printable SolidWorks structures, and design-report preparation. Use when working on this project's K230 Python files, STM32 firmware, mechanical CAD, camera gantry, tests, or report, especially when prior corrections and verified baselines must be preserved.

valentinozegna Skill

Ships finished work instead of a confession. Removes the closing "one honest caveat", "worth flagging", "Say the word and I'll", "Want me to?" that hand a known, fixable defect back to the user as prose. Use on ANY task that ends in a report: code, benchmarks, audits, migrations, refactors, research, writing, data pipelines. Also use on "no caveats", "no-honest-caveat", "just finish it", "don't ask, do it", "stop asking", "close the gaps", or any complaint about permission prompts, handed-back TODOs, or a summary where a result belongs.

vincent-tvincent Skill

Use this project's Abralia keyboard after the user explicitly asks to enable it for this task; report progress and attention through the available Abralia tools.

reganzed Skill

方案类文档与 Markdown 输出的落盘规范——任何时候用户要求"写方案""出方案""给方案""输出 Markdown 文档"时，都必须将方案保存为项目 `Assets/Docs_md/` 目录中的 .md 文件，而不只是在对话中输出。当用户说"写方案""出方案""给方案""做个方案""设计方案""技术方案""实现计划""调研/对比""排查记录""输出成 md""存成 markdown"等，或任何要求产出成文方案 / Markdown 文档的场景时，使用本 skill。

reganzed Bundle

Create clean, professional matplotlib visualizations for technical blog posts and learning documentation. Use when creating plots, charts, or data visualizations for blogs, notes, or papers that need clean styling with English labels. Triggers include "plot", "visualize", "draw chart", "create figure", or when writing technical documentation that needs illustrations.

wutongyuonce Bundle

Manage a non-trivial pull request from issue/spec clarification through design, implementation, review fixes, test ablation, and final re-review. Use whenever implementing an issue across modules, drafting or updating a PR design report, responding to PR review, auditing rule ownership or module seams, or preparing a PR for final review—even when the user only says “continue this PR”, “处理最新 review”, or “检查设计边界”. Skip for tiny one-line fixes and purely read-only PR summaries unless the user asks for design or process analysis.

keyuchen21 Bundle

This skill should be used when the user asks to "help me prepare a talk", "how do I start my talk", "how should I end my talk", "review my slides", "review my deck", "critique my presentation", "be my audience", "let me practice my talk on you", "prepare for my job talk", "prepare for my conference talk", "prepare for my thesis defense", "prepare for my oral exam", "make my talk memorable", or asks what Patrick Winston's MIT lecture "How to Speak" says about speaking. Coaches a talk from empty page to rehearsal using that lecture's rules.

serialexperimentslainnnn Skill

Use when writing, reviewing or productionizing R code - .R/.Rmd/.qmd/.Rproj files, DESCRIPTION, NAMESPACE, renv.lock, .Rprofile, .lintr, _pkgdown.yml, testthat tests, roxygen2 blocks, tidyverse/dplyr/ggplot2 or data.table pipelines, non-standard evaluation with {{ }} and .data, CRAN/Bioconductor/Posit Package Manager repositories, Shiny apps (app.R, server.R, ui.R), Plumber APIs (plumber.R), Quarto or R Markdown reports, Rcpp/cpp11 native code, or rocker/r-base container images.

serialexperimentslainnnn Skill

Use when browser performance must be measured and defended as a budget — Core Web Vitals LCP, INP and CLS thresholds, the web-vitals JavaScript library, Chrome UX Report (CrUX) and the CrUX API, PageSpeed Insights, real user monitoring at the 75th percentile, Lighthouse CI with lighthouserc.js and budget.json performance budgets, a performance budget that fails the build, WebPageTest and its filmstrip and waterfall, DevTools Performance panel traces, Long Animation Frames API, long tasks and main-thread blocking, Total Blocking Time, Speed Index, Time to First Byte, JavaScript bundle cost and hydration cost, layout shift attribution and sources, content-visibility, requestIdleCallback and scheduler.yield, third-party script cost, or justifying performance work with conversion and bounce-rate data.