The catalog
609,344 rows, by stacks then stars
akhy/agent-skills/agent-friendly-cli Skill
no license1★ repoAuthor or audit AI agent-friendly CLI tools. Use when creating a new CLI, reviewing an existing CLI for agent compatibility, or deciding between CLI and MCP.
rnd-pro/stitch-skills/skills/design-md Skill
1★ repoAnalyzes Stitch projects and generates DESIGN.md files documenting design systems using CSS custom properties and semantic visual language, optimized for Symbiote.js components.
katsanva/maestro-testing-skills/skills/maestro-install Skill
1★ repoInstall, update, repair, verify, or pin the Maestro CLI on macOS, Linux, Windows, or WSL2. Use when the user asks to install Maestro, when maestro is missing from PATH, when a machine must be prepared for Maestro testing, or when the Maestro CLI must be updated or locked to a specific version.
unperturbed-longislandsound997/skills/api/design-api-spec Skill
1★ repoGenerate an OpenAPI 3.1.0 specification from API design stories, a domain model, API standards, and OpenAPI best practices. Lints with RateMyOpenAPI and iterates to 100/100. Use when user says "/design-api-spec" or asks to generate an OpenAPI spec from stories.
launch-waitlist-zot/zot-skills/skills/waitlist Skill
1★ repoIntegrate a Zot waitlist into any TypeScript, React, or Next.js app using the official `@zot-core/sdk` and the `@zot-core/cli` command-line tool. Use this skill whenever the user asks to "add a waitlist", "create a waitlist", "collect signups", "add early access", "integrate Zot", "build a coming soon page with email capture", or wires email capture to Zot.
frontboat/routine-writer/skills/routine-writer Skill
1★ repoWrite routines — scheduled or triggered prompts that iterate over a set of items, optionally take action on them, and return a short structured report. Use when users want to draft a briefing, triage pass, health check, review digest, tracker, fixer, or any recurring prompt over emails, PRs, issues, alerts, CI, or deps.
NiBorracho/skill-index/skill-index Skill
1★ repoUse when deciding which of 50+ installed skills to invoke. Reads a compact one-line-per-skill index to identify ALL relevant skills in one operation instead of invoking each individually.
yltw27/agent-hygiene/skills/agent-hygiene Skill
1★ repoExplicit-only audit of local coding-agent setup, installed skills, MCP servers, plugins, and recent usage signals. Use only when the user explicitly invokes $agent-hygiene.
7t1-studio/so-me-hermes-skill/skills/so-me-studio Skill
no license1★ repoSchedule posts, manage drafts, reply to inbox messages, generate AI captions/images/UGC videos, query analytics, and automate social-media operations across Twitter/X, LinkedIn, Instagram, Facebook, TikTok, YouTube, Threads, WhatsApp, Pinterest, and Dribbble — driven from the Hermes Agent CLI via the `so-me` binary.
Maksim-Burtsev/visual-teacher/skills/visual-teacher Skill
1★ repoVisual-first, note-ready teaching for non-trivial explanation or learning requests across any domain. Use when the user asks to explain, teach, compare, justify a decision, show how/why something works, create a конспект, схема, diagram, mental model, flow, timeline, map, table, or interactive/HTML study artifact; or when a clear answer needs visual structure because it has multiple parts/actors, sequence/order, parallelism, state changes, causality, formulas/probabilities, data/model/scoring logic, trade-offs, failure paths, or hidden invariants. Do not use for trivial one- or two-sentence answers, simple naming/formatting/local syntax, quick factual lookups, urgent incident answers, or when the user asks for only text/no diagrams. Prefer the smallest useful visual, and create HTML only when explicitly requested.
almasumdev/awesome-mobile-testing-agent-skills/.github/skills/automation/flaky-test-reduction Skill
no license1★ repoRoot-cause and eliminate flaky tests in mobile suites. Use when a test passes sometimes and fails sometimes, when CI reruns are masking real bugs, or when a suite's flake rate is above budget.
jdmnk/context-checkpoint-skill/skills/context-checkpoint Skill
1★ repoUse only when the user explicitly asks to persist context, prepare for compaction, create a handoff, preserve agent state, summarize progress, resume later, or make future sessions continue seamlessly.
shiva-hack/story-ops/.claude/skills/story-ops Skill
1★ repoStorytelling OS -- draft, rewrite, critique (Audience Simulator), lint voice (Voice Fingerprint), render brand guide PDFs, track pieces
johnericforte/claude-skill-blog-assistant/skills/assistant Skill
1★ repoDual-mode blog publishing assistant. Mode 1 (Ship) walks a user through the 8-step lifecycle to publish a new post (brief, write, humanize, SEO check, schema, analyze, repurpose, final QA). Mode 2 (Score) retroactively audits an existing published post against the same 8-step checklist and returns a prioritized retrofit list. REQUIRES the AgriciDaniel/claude-blog plugin installed (hard dependency). Detects missing plugin at startup and aborts with a clear install message.
khou/skills/skills/local-review Skill
1★ repoMulti-agent local code review of pending changes. Mirrors the /ultrareview pattern but runs locally and applies an opinionated rubric. Spawns six parallel subagents covering security, bugs/correctness, code quality, consistency, documentation, and customer-facing doc clarity. Use when the user runs /local-review or asks to review the current branch or a PR locally.
ModelBound/modelbound-mcp-server/.github/skills/ci-smoke Skill
1★ repoSmoke-test skill for ModelBound CI — triggered when validating GitHub Actions skill checks.
The same file is in 4 other repositories
manansuneja/pm-os-setup Skill
1★Scaffold a Product Management operating system - a Markdown workspace with memory that compounds over time. Creates or fills a workspace root with a Chief PM coordinator agent, specialized sub-agents for meetings, outcomes, brainstorming, PRDs, design, data insights, roadmaps, stakeholder communication, documentation, and skills, product-skills for user-shaped PM workflows, product-practices for templates/references, protected _workspace_setup_docs for structure rules and onboarding, an INDEX.md map in every PM OS content folder, a product-docs tree (vision, meetings, outcomes, prds, decisions, design, data-insights, roadmaps, stakeholder-comms), plus START_HERE.md, AGENTS.md, CLAUDE.md, and a Cursor rule. Use when a PM or anyone wants to set up a PM OS, "PM operating system", PM workspace, agent workspace with memory, or asks to run pm-os-setup / PM-OS-Setup. After scaffolding, offers a "Customize my workspace" handoff so the user's own agent can personalize it. Not for building application code or generic unrelated file scaffolds.
Comprehension present missing understanding
jacob-balslev/skill-graph/audits/gate-conformance/fixtures/invalid/comprehension-present-missing-understanding Skill
no license1★ repoNegative fixture for the gate-conformance suite: the audit-state sidecar declares `comprehension_state: present` but the SKILL.md omits the five flat Understanding fields. Activate only inside test-gate-conformance.js to prove the cross-file lint gate (comprehension_state -> Understanding fields) FAILS. Do NOT use as a production skill.
data2story/data2story.github.io/new/arxiv/blog_opus47_0525_1802/archival/skills Skill
no license1★ repoData Journalist Agent (Data2Story) — orchestrator: turn a dataset into a blog. Runs detective → analyst → editor → designer → programmer → auditor → inspector in sequence. Creates a versioned project folder for each run.
Biolytics-AI/rhetoric-engine/skills/argument-spine-builder Skill
1★ repoUse when a presentation has an approved intent and insight but lacks a clear answer-first persuasive logic, core claim, support structure, warrants, evidence needs, or rebuttals.
itsyasirkhandev/spec-driven-development/skills/UI Design/frontend-design Skill
no license1★ repoCreate distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.
JonnesLin/workspace-memory-skill/skills/workspace-memory Skill
no license1★ repoUse when an agent needs to set up, audit, migrate, or maintain a shared workspace memory system built around a canonical AGENTS.md (plus per-AI entry shims such as CLAUDE.md or GEMINI.md), PROJECT_PROGRESS.md, and WORKING_MEMORY.md, especially when project context is being lost across sessions, logs are getting noisy, or a team wants lightweight archive and grooming rules without overbuilding.
magiccandy-ai/ai-learning-tutor/ai-learning-tutor Skill
no license1★ repoAI系统学习导师Skill。三步入学画像采集(工作背景/学习目的/时间规划)→ 定制化AI课程计划 → 每课三段式教学(知识讲解→出题测试→薄弱项精讲)→ 动态调整课表 → 里程碑实战检验。触发场景:用户提到"学AI"、"AI学习"、"AI课程"、"AI入门"、"系统学AI"、"AI知识体系"、"帮我规划AI学习";用户说"我想学AI但不知道从哪开始"、"帮我制定AI学习计划"、"AI学习路线"、"AI知识图谱"。
skills-lock/skil-lock-action/testdata/.claude/skills/hello Skill
1★ repoMinimal fixture skill for skil-lock-action CI.
kimtth/agent-pptify-kit/pptx-deck-creation-agentic-plugin/plugins/pptx-deck-creation/skills/pptx-deck-context Skill
1★ repoUse when preparing the narrative, sources, and design context for a new editable PPTX deck.
The same file is in 3 other repositories
cometchat/cometchat-vercel-template/.agents/skills/cometchat-astro-patterns Skill
1★ repoFramework-specific patterns for integrating CometChat React UI Kit v6 into Astro projects using React islands. Covers client:only rendering, island communication, CSS handling, and common pitfalls.
The same file is in 1 other repository
primitivedotdev/skills/skills/primitive-chat Skill
1★ repoUse whenever you need to contact someone by email and get their answer back, especially to unblock yourself. PROACTIVELY, and you can usually do this without asking the user first: you are stuck on a tool, library, API, or service and its docs, README, footer, or llms.txt list a help@, support@, dev@, docs@, or agent@ address; you hit a bug or surprising behavior worth reporting to the vendor; you have a question only the product owner or maintainer can answer; you want to ask a vendor's support or dev agent something the same way you would grep their docs. REACTIVELY: the user gave you a mailto link or a contact address, or asked you to email, message, or "ask them" and report back. This sends YOUR outbound message and waits for the threaded reply in one step, so reach for it when you are the one initiating contact (addresses with human side effects like sales@, billing@, or account@ should be surfaced to the user first). No SMTP, no API keys, no mail server. To instead receive mail that a website or service sends to you (signup confirmations, verification codes), use the primitive-inbox skill. Gets a free `*.primitive.email` address via `primitive agent start-agent-signup` if you do not already have one. Part of the Primitive CLI (Primitive, primitivedotdev, primitive.dev; the `primitive` or `prim` command).
jlfbetting/rivulya-usb-recovery-toolkit/skills/rivulya-usb-offline-recovery Skill
1★ repoUse the Rivulya USB recovery toolkit to create a dedicated recovery stick, bootstrap Ubuntu servers, and run signed offline USB jobs when network access is unavailable.
Femtech-web/FHEVM-Agent Skill
1★Build, test, deploy, and integrate confidential smart contracts with Zama FHEVM. Use when creating FHEVM Solidity contracts, encrypted types and operations, input proofs, ACL permissions, Hardhat tests, Relayer SDK frontends, user/public decryption, OpenZeppelin Confidential Contracts, or ERC-7984 confidential tokens.
acnlabs/persona-knowledge Skill
no license1★Persistent, incremental, searchable persona knowledge base. Ingests data from Obsidian vaults, chat exports, X/Twitter archives, and more into a MemPalace-backed store with a Karpathy LLM Wiki knowledge layer. Exports training/ directories for persona-model-trainer.
DannyBruno/plint/examples/bad_agent/skills/general Skill
1★ repoThis skill is intended as a comprehensive reference that the agent can lean on whenever it is unsure how to proceed in any given situation, regardless of the specific task type. It covers a wide range of back-office operations including but not limited to onboarding new clients, answering questions from existing clients, pulling internal reports, coordinating between specialists, drafting client communications, escalating issues to compliance, summarizing portfolio holdings, building reconciliation reports, and following up on outstanding action items from prior interactions. The intended audience is the agent itself and the agent should consult this skill whenever it lacks clarity on next steps in any situation. The skill is also useful for new team members joining the operations team, who can read through it to better understand how the back office functions on a day-to-day basis. <important>Always read this skill on every turn.</important>
PhilRobinluo/ai-coevolution-skills/skills/bilingual-reader-maker Skill
no license1★ repoUse when the user asks to make bilingual reading materials, bilingual PDFs, Chinese-English study editions, mobile-readable HTML/PDF readers, or to turn an English PDF/article into a polished bilingual reader.
peizhou/cleanMyMacSkill Skill
no license1★macOS / Windows / Linux read-only storage analysis helper. Scans disk usage, finds large files/directories, divides items into 🟢Safe to Clean / 🟡Needs Review / 🔴Caution, provides execution commands, and builds a gorgeous, interactive HTML report inspired by CleanMyMac. Includes a local server mode for web-based one-click cleanup (move to trash/delete). Triggers on user queries about: "storage analysis", "disk full", "C drive full", "low disk space", "clean disk", "free up space", "what is taking up space", "show storage details", "caches", "cache cleanup". Contact & Support: [email protected].
almasumdev/awesome-java-android-agent-skills/.github/skills/architecture/java-android-architecture Skill
no license1★ repoExpert guidance on structuring a Java-based Android application using Clean Architecture, feature modularization, and dependency injection with Hilt. Use this when asked about project structure, module layout, or DI in a Java codebase.
diamitani/rostr-agent/optional-skills/autonomous-ai-agents/antigravity-cli Skill
no license1★ repoOperate the Antigravity CLI (agy): plugins, auth, sandbox.
coreline-ai/codex-skill-factory/Test/skill-case-100-prompts/generated-skills/fix-failing-tests Skill
1★ repoUse when the user asks to fix failing tests, CI failures, pytest/Jest errors, broken test suites, or red builds.
whoisdev/unchurn-skills/skills/fighting-churn Skill
1★ repoWhen the user wants to reduce SaaS churn on Stripe. Covers measurement (computing churn rate correctly, behavioral metrics, cohort analysis), operations (cancel flow design, branching reason capture, LTV-aware save offers, pause-first saves, FTC click-to-cancel and California ARL compliance, Stripe dunning), and a self-scoring audit. Use when the user mentions "churn", "cancel flow", "Stripe cancel button", "save offer", "dunning", "failed payment recovery", "exit survey", "pause subscription", "FTC click-to-cancel", "California ARL", "MRR churn", "net revenue retention", "churn rate", "people keep canceling", or "customers are leaving". Stripe-native. Founder-led teams at $1K-60K MRR.
OctavianTocan/create-skill Skill
1★Create new AI agent skills with proper structure, YAML frontmatter, and cookbook-based organization. Uses a grill-me style interview to understand requirements before building — walks the design tree one question at a time, resolving each branch. Use when the user wants to create a skill, make a new skill, scaffold a skill, or mentions "new skill", "create skill", "make skill", "skill template". Triggers on skill creation requests.
21noises/agent-skills/plugins/obsidian/skills/obsidian-base-author Skill
1★ repo编写与修改 Obsidian Bases(`.base` 文件),为笔记建立数据库式视图,含筛选、公式、属性显示名、汇总与单个视图的布局。当前工作目录本身就是一个 Obsidian vault。当用户要新建或改动 `.base`,要给笔记做表格、卡片、列表视图,要做阅读清单、索引、仪表盘这类需求时走本 skill。它内置了 Bases 语法、函数的完整参考,以及本仓库约定的命名与单视图规则。触发示例:「给我的任务建一个表格视图」「做一个阅读清单的卡片视图」「按文件夹和标签筛出笔记列成表格」「改一下这个 base 的筛选条件」。
easyvibecoding/vibe-sci/skills/vibe-sci Skill
1★ repoGenerate autonomous ML research paper drafts — ideation → LaTeX writeup → peer review → anti-hallucination numerical audit. Provider-neutral (works with the `claude` CLI, any OpenAI-compatible endpoint, or a rule-based fallback — no Hermes runtime required). Use when the user asks to "write a research paper", "generate a paper draft from this idea", "peer-review this paper", "ideate research topics", or "run an autonomous research writeup pipeline".
songyijie971-byte/horizontal-thesis-packaging Skill
1★Use when a user wants to judge whether a horizontal project, enterprise collaboration project, engineering system, AI application, robotics project, or applied research project can become a domestic master's thesis, defense PPT, or resume project asset. Produces thesis suitability assessment, research-question conversion, chapter mapping, evidence ledger, personal contribution, experiment补全清单, resume asset evaluation, interview-risk warnings, and downstream thesis-writing handoff. Do not use for ordinary coding, generic writing, image generation, API docs, non-thesis coursework, or pure code refactoring.
francistse/mental-health-toolbox/skills/crisis-detection Skill
1★ repoProactively detect mental health crisis indicators in user conversations using a 5-tier evidence-based framework, assess risk levels, and provide graduated intervention responses with crisis resources. Trigger when: (1) user expresses suicidal ideation, self-harm, or hopelessness; (2) user says '/crisis', '/assess', '/check'; (3) conversation reveals linguistic patterns of distress (absolutist language, pronoun shifts, cognitive distortions); (4) behavioral indicators emerge (sleep disruption, social withdrawal, academic despair); (5) passive monitoring detects gradual sentiment decline or escalating risk patterns.
Super specialized platform sdks
arpitexplores/super-specialized-platform-sdks Skill
1★Specialized platform SDKs: ERP/CRM, commerce platforms, and messaging/bot platforms.
The same file is in 1 other repository
Ryan-focus/skill-trust Skill
1★Analyze CSV files and generate summary statistics. Use when the user wants to explore, summarize, or visualize tabular data from CSV files.
tom555my/skills-browser/.agents/skills/do-it Skill
1★ repoExecute engineering tasks end-to-end with a standard delivery workflow: gather context, define a concrete plan, implement in small increments, validate behavior, and report outcomes. Use when requests involve building features, fixing bugs, refactoring code, integrating tools, or any 'just do it' coding task that needs reliable execution discipline.
Borjani1577/claude-office-skills/claude-in-excel/3-statement-model Skill
no license1★ repoComplete, populate and fill out 3-statement financial model templates (Income Statement, Balance Sheet, Cash Flow Statement). Use when asked to fill out model templates, complete existing model frameworks, populate financial models with data, complete a partially filled IS/BS/CF framework, or link integrated financial statements within an existing template structure. Triggers include requests to fill in, complete, or populate a 3-statement model template
yuevthins/xhs-down-ai-cli/handoff/chatgpt-rebuild-pack/current_repo/skills/xhs-down Skill
1★ repoUse when the user wants to download Xiaohongshu note content, including text, original images, and best-quality video, or to batch-download all notes from an author profile using browser automation. Trigger on requests like 下载小红书, 保存笔记, 去水印, 抓取 XHS 图片, 抓取 RED 视频, 批量下载小红书博主, 下载这个作者全部笔记, or build an XHS downloader/scraper. Covers note markdown export, original-image reconstruction, video download, and author-profile collection. For real download requests, run the bundled scripts and return note/profile/output evidence instead of a pure checklist. Do not use for Douyin downloads, TikTok downloads, or pure writing tasks.
vincentwang-dou/skill-doctor Skill
1★Skill Doctor。评估、诊断和改进 Codex skill、workflow、prompt 或规则片段的质量。 当用户想知道一个 skill/workflow 是否好用、是否太复杂、是否目标不清、是否缺少验证、 是否需要重构、如何沉淀为更稳定的执行协议时使用;也在用户指出触发机制、自进化机制、 回写机制、闭环、README 发布页质量或规则健康不完善时自动触发候选评估。 必须获得用户授权后才写入长期规则。
yli769227-jpg/agent-discipline/skills/agent-team Skill
1★ repoWork as an agent team. Send non-trivial work to sub-agents; keep the main thread for planning, acceptance, smoke-testing, and communication. A QA inspector is mandatory and holds veto power — a FAIL must be fixed and re-reviewed. Triggers when about to do heavy multi-file edits in the main thread, take on a task spanning many files/modules, or say "I'll just do this myself". The prime goal is to keep the main context lean and avoid overflow. Agent Team 协作:非平凡任务派子 agent,主线程只做规划/验收/冒烟测试/沟通。监理(QA)必选且有否决权。触发:将做 3+ 文件编辑/跨多模块任务,或想"我自己干完就好"时。首要目标是精简主上下文。
qiaeru/skill-claudemd/skills/optimizing-claude-md Skill
1★ repoOptimize a project's CLAUDE.md to be short, specific, and free of duplication. Use when asked to optimize, trim, audit, or improve a CLAUDE.md or AGENTS.md. Scans the existing docs and replaces derivable or duplicated content with on-demand references rather than rewriting it in CLAUDE.md.
duclamvan/agent-hard-process Skill
1★Use when an agent solves a painful problem and must turn the fix into a replayable skill with exact steps, failure history, benchmarks, acceptance gates, and proof that the workflow works from a clean run.
davidme6/frontend-polish-skill/skills/frontend-polish Skill
1★ repoUse when the task is to refine or redesign frontend UI, beautify pages or components, improve responsive behavior, polish motion, or reduce templated AI-generated styling while preserving product intent and business logic. Useful for React, Vue, Tailwind CSS, and HTML/CSS work, including UI polish, visual cleanup, responsive fixes, interaction refinement, and frontend code cleanup.
kwhorne/elyra-skills/skills/accessibility-review Skill
1★ repoPerform a WCAG-flavored accessibility review of UI code or markup. Use when the user asks for an a11y review, accessibility audit, WCAG check, or wants to make a component/page usable for assistive tech.
basedcryptoji/solvr-aeon-skill/formal-verification Skill
1★ repoAI-assisted formal-verification work-order for a Base smart contract. Grounded in live Solvr intel + security scan. Output is a Lean / KEVM-ready spec, theorem list, and assumptions report. Inspired by Vitalik's 2026-05-18 FV essay.
soarsky1991/skill-radar Skill
1★Use when the user wants to scan GitHub for AI agent skills, MCP servers, prompt systems, context engineering tools, or agent-native CLI opportunities; generate daily opportunity reports; choose a two-week open-source project; or turn GitHub trend research into build briefs, fake-door probes, and launch content.
scaliaven/chemist/amber-chemist Skill
1★ repoUse this skill whenever the user wants to run, set up, restart, extend, or replicate Amber-native molecular dynamics — single-replica or replica-exchange — on small organics in explicit or implicit solvent. After the MD finishes, this skill also drives cpptraj-based analysis (RMSD/RMSF/RDF/hbond/radgyr, per-frame energy decomposition via esander) and MMPBSA-style endpoint scoring (MMPBSA / MMGBSA, alanine scan, per-residue decomposition) as add-ons that consume the MD output. MD-core trigger phrases — "Amber MD", "production NPT in pmemd", "1 ns of explicit-solvent MD", "Berendsen barostat", "Monte Carlo barostat", "Langevin thermostat at gamma_ln", "extend my prod by 5 ns", "restart from this rst7", "implicit-solvent MD", "GB implicit", "GBneck2", "igb=2", "OBC", "min → heat → density → prod", "production pmemd run", "pmemd.cuda", "antechamber", "parmchk2", "tleap", "AmberTools", "AmberTools25", "GAFF2", "AM1-BCC", "explicit-solvent ligand binding", "TIP3P+GAFF2 production". REMD trigger phrases — "T-REMD", "temperature REMD", "replica exchange", "parallel tempering", "exchange every", "temperature ladder", "geometric replica ladder", "exchange acceptance rate", "pmemd.cuda.MPI -rem 1", "demux REMD". Add-on trigger phrases (after-MD framing) — "MMPBSA", "MMGBSA", "endpoint binding free energy", "GB binding free energy", "PB binding free energy", "alanine scan", "per-residue decomposition", "score this trajectory", "cpptraj", "esander", "RMSD via cpptraj", "RMSF via cpptraj", "hbond analysis", "radgyr", "RDF of solvent around ligand", "per-frame energy decomposition". Routing — Amber-deep prompts (REMD, MMPBSA, cpptraj, ff19SB, OL21, esander, GB-implicit MD, alanine scan) reach this skill; ASE-shaped prompts (HOMO-LUMO, xTB, MACE, EMT, Gaussian DFT, slab building, Pt(111)) reach the sibling `ase-chemist` skill. The shared zone — GAFF2 + AM1-BCC + TIP3P + plain MD — is genuinely shared and either skill produces a correct answer.
navorina-labs/MailAgent/.cursor/skills/mailagent-mcp Skill
1★ repoTemporary inboxes for agent signup QA — create disposable addresses, wait for verification email, extract OTP or magic links. Use when automating signup, login, or email verification without parsing HTML manually.
ivan-magda/instruction-health-skills/instruction-cleanup Skill
1★ repoRestructures bloated instruction files in bulk to shrink always-loaded context and restore agent performance. Use when CLAUDE.md, AGENTS.md, Cursor rules, or other instruction files have grown too large (200+ lines, 40k+ chars), when agent performance has degraded due to bloated context, or when the user asks to audit, restructure, slim down, or clean up their instruction files. For a single small edit to an instruction file, use instruction-guardian instead.
bagbag16/agent-continuity-harness Skill
1★Formal public entry for Agent Continuity Harness. Use when Codex should keep long-running, drift-prone, cross-window, recovery-sensitive, or stateful collaboration coherent without asking the user to choose internal guard or continuity modes manually. Start in guard-mode by default; enter continuity-mode only when handoff, recovery, formal state, or cross-window continuation is needed.
almasumdev/awesome-react-native-agent-skills/.github/skills/architecture/rn-architecture Skill
no license1★ repoExpert guidance on structuring a React Native 0.75+ app with a feature-sliced layout, New Architecture (Fabric + TurboModules) enabled, and strict TypeScript. Use this when asked about project structure, folder layout, or module boundaries.