agentsclimarketplace

Engineering

98,276 rows, by stacks then stars

  • Phere 9 agent manager

    BAKUGOS1/SkilledAgents-Toolkit/plugins/phere/skills/phere-9-agent-manager Skill

    0★ repo

    Use when managing the Phere/Shaadikhata app through the named Phere 9-Agent Manager developer workflow, including enterprise-style engineering agents, planner, product, frontend, backend, Supabase/database, tester, production, security, architecture, performance, observability, documentation, data/AI, truthfulness checks, and optional post-code business agents such as marketing and sales. Trigger when the user asks for agents, an agent manager, deep agent training, planner/tester/production agents, evidence-based app maintenance, release coordination, or a multi-agent development team.

  • My backend conventions

    AGilbertDev/agilbertdev-recipes/skills/my-backend-conventions Skill

    no license0★ repo

    My reusable coding-agent conventions and a curated skill set, pulled into any project like a versioned package.

  • Flutter clean architecture

    Mkhira/flutter-clean-architecture/skills/flutter-clean-architecture Skill

    0★ repo

    Use this skill when creating a new Flutter project, implementing Flutter or Dart features, writing code in an existing Flutter project, refactoring Flutter/Dart code, or reviewing Flutter code. Trigger whenever the user asks to create a Flutter project/app by name, or when the current workspace is a Flutter project and the task involves implementation, coding, API integration, models, Bloc/Cubit, UI, theming, localization, dependency injection, networking, tests, or code review — even if the user does not explicitly say "clean architecture". For new project creation, ask the user for the directory and project name before running flutter create.

  • Simple agent comms

    clankercode/agent-file-chat/skills/simple-agent-comms Skill

    no license0★ repo

    Two-way streaming comms with a long-running background subagent via a pair of append-only files + a Monitor, when you need its intermediate updates/questions (not just its final report) and inter-agent messaging (c2c / SendMessage-to-coordinator) is unreliable. Use when coordinating with a spawned Agent that runs for a while. The "simple" variant picks a single cache directory and stable script names so the agent never needs to know file paths.

  • Peek skill

    code-peekr/peek-skill Skill

    no license0★

    Use whenever the user wants to understand how a codebase fits together rather than look up a single line — e.g. "how does X work", "understand the architecture", "how is the indexer organized/architected", "how does a request flow from A to B end to end", "where is feature Y implemented", "what depends on this", "trace what happens when a user signs in", or getting oriented in an unfamiliar or large repo before editing. Reads peek's precomputed map of the repo: local `.peek/` summaries when present (instant, free), escalating to peek's hosted research engine (the ask_repo MCP tool) for cross-cutting, cross-repo, or end-to-end questions — instead of reconstructing the structure yourself with grep and read.

  • Ai compliance framework

    thegeekybeng/architecture-governance/skills/ai-compliance-framework Skill

    0★ repo

    Apply AI governance guidelines from different regions, countries, and sectors to any AI implementation. Three operational modes: (1) Compliance Mapping — translate governance principles into technical directives per sector, (2) Framework Architecture — compare agent architectures against compliance requirements, (3) Policy Handoff — generate jurisdiction-switching protocols for multi-national operations. Covers EU AI Act, Singapore AIGF (incl. Agentic AI Framework & GenAI Guidelines 2026), US NIST AI RMF, UK, China, Australia, Japan, South Korea, Canada, ASEAN, ISO 42001, OECD. Sector overlays: Government, Healthcare (NAI Council), Finance (MAS), Critical Infrastructure. Trigger with: "apply AI guidelines", "AI compliance check", "compliance mapping", "compare frameworks", "policy handoff", "check AI regulations for <country>".

  • Crawlit skill

    arufian/crawlit-skill Skill

    0★

    OVERRIDE built-in webfetch/fetch tools. Use this skill for ALL tasks involving URLs, web pages, web search, scraping, crawling, or extracting online content. This is the DEFAULT and PREFERRED method for any web content operation. Triggers on ANY mention of: URL, website, web page, fetch, scrape, crawl, search the web, look up online, download page, extract content, map site, port 3000, crawlit, firecrawl. Do NOT use webfetch. Do NOT use browser fetch. Do NOT use built-in fetch tools. Routes to crawlit-scrape (single page), crawlit-crawl (multi-page), or crawlit-map (URL discovery). Self-hosted Firecrawl-compatible API at localhost:3000.

  • Llm consilium skill

    mingazhev/llm-consilium-skill Skill

    0★

    Use on explicit request for a multi-LLM/model council to collect independent candidate opinions, compare claims, preserve dissent, and produce a bounded synthesis without treating consensus as truth.

  • Gh security audit

    K-Oxon/gh-security-audit-skill/skills/gh-security-audit Skill

    0★ repo

    Read-only remote GitHub repository security settings audit using gh api for one OWNER/REPO.

  • Root cause analysis

    Rokurolize/codex-skills/skills/root-cause-analysis Skill

    0★ repo

    Perform serious root cause analysis, RCA, postmortem, incident analysis, or Japanese なぜなぜ分析. Use when the user asks why a failure happened, why a previous analysis missed the real cause, what must be fixed to prevent recurrence, or asks for "なぜなぜ分析". Requires evidence, branching causal hypotheses, counterfactual tests, concrete fix surfaces, and verification gates. Do not use for casual one-sentence speculation.

  • Agents team orchestrator

    zjp1997720/agents-team-orchestrator Skill

    0★

    Orchestrate long, complex Codex tasks with coordinated sub-agents, shared local context, a machine-readable state ledger, task boards, packet specs, approval gates, handoff files, and final integration discipline. Use when the user explicitly asks for Agents Team, sub-agents, delegation, parallel agent work, expert simulation, or when a long-running task needs durable coordination across research, writing, coding, review, or document production.

  • Logging best practices

    Turbovadim/turbo-skills/skills/logging-best-practices Skill

    no license0★ repo

    Design, implement, migrate, or review production logging and observability using wide events/canonical log lines, structured high-cardinality and high-dimensionality telemetry, query-driven debugging, OpenTelemetry enrichment, and tail sampling. Use when Codex is asked to add or improve logs, audit logging quality, design log schemas, instrument request lifecycles, correlate events across services, reduce noisy log lines, debug why logs are not useful, or review code for logging and observability correctness.

  • Android security checklist

    esaldgut/ai-native-engineering-workspace/global-skills/android/android-security-checklist Skill

    0★ repo

    Systematic Android security checklist aligned with OWASP MASVS and platform guidance — secrets via DataStore + Google Tink (AeadSerializer) + Android Keystore (NOT the deprecated EncryptedSharedPreferences), TLS cert pinning via Network Security Config (SPKI hash) and/or OkHttp CertificatePinner with a backup pin, OAuth via Chrome Custom Tabs + PKCE + AppAuth-Android (WebView is forbidden), and manifest hardening (android:exported, allowBackup=false, usesCleartextTraffic=false). Use when handling tokens, network security, sign-in, or attack-surface review on Android.

  • Port skill collection

    vcolombo/port-skill-collection/skills/port-skill-collection Skill

    0★ repo

    Port a Claude Code plugin or external skill collection (a GitHub repo of SKILL.md directories) into Hermes Agent. Use when installing, importing, adapting, refreshing, or auditing existing skills from Claude Code, Codex, Cursor, OpenCode, or agentskills-style repos — e.g. "port this skill repo into Hermes", "add the skills from github.com/x/y", "refresh the ported skills from upstream", or "audit this skill collection before install". For authoring new skills from scratch, use skill-creator instead; this skill migrates existing ones. Also use when publishing a ported or locally maintained skill — "publish this skill to the Hermes Skill Hub / HermesHub", "create a private repo for my custom skills", or "add CI/release safety gates to my public skill repo".

  • Agent skill instruction creation

    Nick2bad4u/codex-skills/skills/agent-skill-instruction-creation Skill

    no license0★ repo

    Creates agent skill and instruction surfaces. Use when authoring SKILL.md folders, AGENTS.md/AGENTS.override.md, CLAUDE.md, Copilot or Cursor rules, or when asked to bootstrap, design, scaffold, or migrate reusable agent guidance.

  • Agent never stop

    Hommchen/agent-never-stop Skill

    no license0★

    Agent Never Stop — Frontend autonomous polling loop, reasoning → wait → reasoning → wait, never stops

  • Agent

    teejayai/Omni-design-skill/omni-design-skills/agent Skill

    no license0★ repo

    OmniONE — AGENT product design skill. Mobile field-sales-agent app. Gamified, encouraging, card-driven, TEAL primary action, soft radii (5/10/20). Extends the shared omni-design-system base.

  • Code simplifier

    yidasanqian/codex-code-simplifier/plugins/code-simplifier/skills/code-simplifier Skill

    0★ repo

    Use when Codex should simplify, refactor, or clean up recently modified code while preserving exact behavior, especially after an implementation session, before commit, or when code has become overly complex, duplicated, deeply nested, or inconsistent with project conventions.

  • Build work memory

    stardustai/stardust-skills/skills/build-work-memory Skill

    0★ repo

    Guided first-run workflow for importing a user's existing work context into Friday Memory from local files, DingTalk documents/wiki, DingTalk AI minutes, and selected DingTalk data through dws. Use when the user wants to initialize Memory, connect work context, import DingTalk knowledge, seed Memory from documents or meetings, or prepare Memory for daily work.

  • Postgres pro

    AojdevStudio/development-kit/.claude/skills/postgres-pro Skill

    no license0★ repo

    Use when optimizing PostgreSQL queries, configuring replication, or implementing advanced database features. Invoke for EXPLAIN analysis, JSONB operations, extension usage, VACUUM tuning, performance monitoring.

    The same file is in 1 other repository

  • Escalation handling

    joyboy257/usk/spec/examples/escalation-handling Skill

    0★ repo

    Universal Skills Library — a Rust schema, CLI, adapter system, and optional registry for portable AI-agent skills across Claude Code, Codex, and future harnesses.

  • Improve codebase architecture

    lyra-engine/lyra-step-2.5-architecture-deepening/skills/improve-codebase-architecture Skill

    0★ repo

    Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable. Deutsch: "architektur verbessern", "refactoring-chancen finden", "deep modules", "code aufräumen", "testbarer machen".

  • Llm pair

    iotashan-llc/llm-pair Skill

    0★

    Use for ALL pairing with peer LLMs. Triggers: the user says "pair with codex", "pair with gemini", "pair with the LLM", "llm-pair", "three-way", or "co-develop with a second model" — OR proactively during substantial defined task/ticket work at the plan, work-item-review, and blocker boundaries. Runs up to TWO read-only peers — Codex / GPT-5.x (`codex exec`) and Google Gemini (`agy`, the Antigravity CLI) — calibrating each peer's model + reasoning effort to the task, and fans out to BOTH for complex work (planning + `big`) so the coordinating Claude reconciles a three-way consensus. Redacts sensitive data to stable placeholders before anything reaches a third-party peer. Self-throttles via a `skip` verdict on trivial work. Falls back to previous-gen Opus only when BOTH external peers are unavailable.

  • Ux workflow check

    sergiobuilds/ux-workflow-check Skill

    0★

    Use when you need to find what a real user would need but nobody wrote down — surfacing tacit assumptions, missing screens/features, and broken workflow links grounded in the user's actual intent, not imagined "best practices." Triggers: "UX check", "find what's missing", "workflow gaps", "is this annoying to use", "review this product for users", or when you're about to declare a one-pass review done. Works for UI, CLIs, reports, and docs — anything a user consumes, at plan stage (a spec) or audit stage (a built product). Not for visual polish (use a design-review tool) or idea/demand validation (use an idea-validation tool).

  • Savepoint workflow

    NyXXiR/savepoint-workflow-skill/savepoint-workflow Skill

    0★ repo

    TDD-first checklist execution with repository-backed savepoints for continuing implementation work across AI agents, accounts, sessions, and providers such as Codex, Claude, Gemini, opencode, Hermes, and other coding agents. Use when starting non-trivial implementation work, bug fixes, refactors, long-running or multi-phase code changes, tasks likely to exceed context or token budget, provider/model/account switching, interrupted sessions, checkpointed progress, resumable implementation, task savepoints, implementation checklists, handoff notes, or another AI/provider continuing the work.

  • Purpletoadmail skill

    hixistudio/purpletoadmail-skill Skill

    0★

    A guide for using PurpleToad Mail with AI agents. Use whenever the user wants to send, receive, search, schedule, or manage email, set up domains/mailboxes/aliases, configure the PurpleToad Mail MCP server, troubleshoot deliverability issues, or integrate email into an app or workflow. Covers installation, authentication, tool selection, common workflows, and guardrails.

  • Agent payments x402

    plagtech/agent-payments-x402 Skill

    no license0★

    Batch payments for AI agents across Base and additional EVM chains, plus payroll, invoicing, and escrow on Base. Send USDC to many recipients in one transaction, run stablecoin payroll, create payment-gated invoices, and lock funds in on-chain escrow. Pay-per-call via x402 — no API keys, no accounts. Powered by Spraay Protocol.

  • Agent verify

    kmloureiro/agent-verify Skill

    0★

    This skill should be used when building or changing a feature in an app, API, or site and you want to verify it works WITHOUT burning tokens on browser automation, screenshots, or DOM dumps. Use it to generate a token-minimal CLI verification harness alongside each feature, run checks in a sandbox, and surface only PASS/FAIL + the failing line. Triggers on "verify this works", "test this feature", "did the change break anything", "check the endpoint", "add tests", or any agent self-verification loop. Browser/visual checks are the fallback of last resort, only for genuinely visual UI, and only at the end.

  • Long task orchestrator

    jerrynullmmo/long-task-orchestrator/long-task-orchestrator Skill

    0★ repo

    长任务编排技能。用于耗时、多阶段、上下文重、容易漂移,或需要子代理、Codex 新线程、命令任务、交接文件、任务包、机器回执、监控文件、闸门放行的任务。分派前先用本技能,让主线程保持轻量编排,并让每个执行单元都有边界清晰的任务包和可验证回执。

  • Hermes skill local web security scan

    web3blind/hermes-skill-local-web-security-scan Skill

    no license0★

    Use when the user wants an authorized local-only active security scan of his own website or web app without third-party APIs, without adding libraries to the target project, and with the target passed as an argument.

  • Ci pipeline

    HM-IT-CODE/forge-skills/skills/devops/ci-pipeline Skill

    0★ repo

    Scaffolds a GitHub Actions CI pipeline tailored to the project's stack. Use when the user says "add CI", "set up GitHub Actions", "add a pipeline", "automate tests on push", or wants lint/test/build to run on every PR. Produces a cached, matrix-aware workflow.

  • Frago source install

    tsaijamey/frago-source-install/skills/frago-source-install Skill

    0★ repo

    This skill should be used when a user wants to install or configure frago from source on their machine (macOS/Linux/Windows) using their existing Claude Code. It guides the agent to clone the repository, build the environment with uv, run the frago server once so product code auto-deploys hooks and runtime config, then verify the artifacts — the user never types a frago command themselves. Trigger phrases: "install frago", "安装 frago", "配置 frago", "frago 源码安装", "set up frago", "frago installation", "deploy frago hooks", "从源码安装 frago".

  • Circleci cli

    hideokamoto/circleci-cli-skills/skills/circleci-cli Skill

    no license0★ repo

    Operate the CircleCI CLI (`circleci` command) safely and correctly — auth/token checks, org slug (UUID) lookup, project create/list, and config-source questions. Use this skill whenever a task involves the `circleci` CLI: checking whether the CLI is authenticated, resolving an org slug or org UUID, creating or listing a project, running `circleci diagnostic` / `circleci info org` / `circleci project ...`, or answering whether CircleCI can `source` a YAML config from an external repo. Also trigger for Japanese phrasings such as 「CircleCI CLI で〜」「circleci の 認証を確認」「プロジェクトを作成」「org の UUID を取りたい」「config を外部参照したい」. This skill exists to prevent specific recurring CLI mistakes; consult it before running any `circleci` command, even when the request looks like a one-liner.

  • Browser automation

    lensetek/Digital-Marketing-Agent_Skills/capabilities/browser-automation Skill

    0★ repo

    Digital-Marketing-Agent_Skills

  • Sd prompt forge

    Edemckb/sd-prompt-forge/skills/sd-prompt-forge Skill

    0★ repo

    Create production-ready Stable Diffusion prompts for stylized character art and multi-character scenes. Use when working with SDXL, Illustrious XL, Pony, Forge or A1111, LoRA stacks, negative prompts, regional prompting, Forge Couple, ControlNet, IP-Adapter, or character-consistency workflows.

  • Goalscout agent

    AKADevelopers/GoalScout-Agent/skills/goalscout-agent Skill

    0★ repo

    Use when the user wants GoalScout Agent help for football or soccer match information, favorite-team setup, live score monitoring, goal alerts, match reminders, player/team tracking, World Cup or league updates, or configuring Codex, Claude Code, OpenCode, Pi Coding Agent, OpenClaw, Hermes, Cursor, or GitHub Copilot CLI football notifications.

  • Agent skill productizer

    wrhy66/agent-skill-productizer/agent-skill-productizer Skill

    0★ repo

    Turn repeatable prompts, workflows, SOPs, automations, or niche expertise into publishable AI agent skills. Use when Codex needs to design, create, audit, package, or launch a GitHub-ready skill with SKILL.md, progressive-disclosure references, reusable scripts/assets, README positioning, launch copy, and monetization paths.

  • Overleaf leaflink sync

    MuQY1818/overleaf-leaflink-sync Skill

    no license0★

    Sync a LaTeX project to an Overleaf project with LeafLink when the user wants to avoid Overleaf Git, browser login, or manual zip upload. Use when setting up or running a LeafLink-based Overleaf sync workflow, including conda environment setup, browser login, clone/pull/push, dry-run checks, or packaging a clean Overleaf source tree from an existing manuscript repo.

  • Diagnose

    vincendiary/agents/skills/diagnose Skill

    no license0★ repo

    Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, or describes a performance regression.

    The same file is in 21 other repositories

  • Audit the oracle coverage

    XyraSinclair/ideonomy/skills/audit-the-oracle-coverage Skill

    0★ repo

    The dangerous case is not "no oracle" — it is a WEAK oracle you falsely believe is exact. Before trusting any test suite, metric, benchmark, eval, or proof as if it settled the question, state explicitly what it CERTIFIES and what it does NOT — the gap between the proxy and the target. The failure it prevents: false certainty from a passing check that silently misses the spec (tests green but the spec was wrong), proxies value (offline metric up, user value flat), or has drifted (backtest on a system that has since changed). Triggers whenever a check comes back positive and you are tempted to call the thing "verified", and as the mandatory middle step between build-the-oracle-before-the-answer and triangulate-without-oracle.

  • Hermes skill prompt

    web3blind/hermes-skill-prompt Skill

    no license0★

    Оптимизация промптов для LLM (/prompt <запрос>)

  • Self learning agent skill

    ardha27/self-learning-agent-skill Skill

    0★

    Use when building an LLM agent that makes repeated decisions with measurable outcomes (trading, deploys, content picks, lead scoring, moderation) and should improve from its own results — especially when the agent repeats past mistakes, ignores what worked before, or its prompt grows unbounded with history.

  • Xray xhttp relay deploy

    superchaospc/xray-xhttp-relay-deploy Skill

    0★

    Install and remotely operate the user's xray-xhttp-relay one-click VLESS+XHTTP+REALITY relay script (superchaospc/xray-xhttp-relay, xray_deploy.sh) on a VPS over SSH. Use whenever the user wants to deploy or manage an XHTTP-transport VLESS/REALITY relay or 中转 — phrases like "装个 xhttp 中转", "xhttp reality 节点", "部署 xhttp 落地/直连", "给 vps 加几个 xhttp 直连/住宅节点", "xhttp 订阅/链接". Covers add/delete/rename nodes (residential SOCKS5 or VPS direct, single or batch), status/traffic/diagnostics, change port, update/restart Xray, monitoring/email alerts, and uninstall. This script transports VLESS over **XHTTP + REALITY** (requires Xray core ≥ 24.10.31), NOT TCP+XTLS-Vision — for the Vision variant use the `xray-relay-deploy` skill instead. Drives the script's interactive 16-item menu non-interactively via SSH; it does NOT reimplement the relay logic.

  • Api security audit

    vishnujchandran/.agents/skills/api-security-audit Skill

    no license0★ repo

    Ultimate API security audit and hardening skill for REST APIs, GraphQL APIs, and webhook endpoints. Use when user asks for API security review, authN/authZ fixes, schema/input validation, abuse prevention, CI security gates, or pre-deploy API hardening.

  • Skill atlas

    luongs3/skill-atlas Skill

    0★

    Find the right public AI-agent skill for a job — and know whether to trust it. Load when about to start a task type (Upwork freelancing, technical interviews, office documents, MCP/tool building, prompt engineering, web/frontend, data analysis, learning English) and you want to know which existing public skills to pull in, rated by source reputation and freshness. Answers "which skill do I load for X, and can I trust it?"

  • Workflow router skill

    Moonweave-Systems/workflow-router-skill Skill

    0★

    Use when a request is broad, ambiguous, or multi-step and the right workflow (implementation, debugging, review, GitHub, research, workflow design, system triage, artifacts) must be chosen before execution.

  • Customer audit

    Evil-Bane/argus/skills/customer-audit Skill

    0★ repo

    Universal, multi-agent, VISION-VERIFIED audit that makes any app, website, tool, CLI, API, or MCP server genuinely great. Fans out one agent per surface and exercises EVERY single control (button, field, toggle, menu, tab, swipe, long-press, drag) for real — screenshotting before/after and looking at the result — across every lens: does each control actually work + persist + sync, the "does the UI lie?" trust audit, UI/interaction craft & motion, fresh-eyes customer journeys, competitive teardown, accessibility, performance, security, microcopy, empty/error/edge states. Adversarially verifies every finding, then returns a ranked fix list (or auto-fixes). Detects and uses the available MCP servers (e.g. Playwright for websites) and routes UI findings through the user's installed design skills. Trigger when the user says "audit", "customer audit", "review every page/screen", "find everything wrong or missing", "test every button", "QA this", "make this badass/premium", "is everything working and in sync", or runs /customer-audit.

  • Autoresearch skill

    agusmdev/autoresearch-skill Skill

    no license0★

    Run an iterative, measured optimization loop — try an idea, benchmark it, keep what improves the metric, revert what doesn't, repeat — leaving an auditable trail of commits. Use this when the user wants to ITERATIVELY optimize a measurable target and expects to try many attempts: speeding up a test suite or build, shrinking a bundle/binary, cutting latency or memory, tuning hyperparameters to lower a loss, raising a Lighthouse/benchmark score. Strong fit when the metric is NOISY (needs repeated runs to trust), correctness must be preserved while optimizing, or a reviewable record of what was tried and kept matters. Trigger for "run autoresearch", "optimize X in a loop", "set up experiments to make X faster", "keep trying changes until the benchmark improves", "tune this until the loss is low", or "measure it properly, don't guess". Less suited to one-shot fixes you already know — see the "When NOT to use this" note before reaching for it on trivial changes.

  • Agent pr review

    ikhattab/agent-pr-review/skills/agent-pr-review Skill

    0★ repo

    Reviews agent-generated pull requests for CI gaming, code duplication, hallucinated correctness, agent abandonment, and workflow security risks. Use when reviewing pull requests authored by AI agents, Copilot, Cursor agents, or coding bots; when the user asks to review an agent PR, AI-generated code change, or automated pull request; or when assessing whether an agent PR is safe to merge.

  • Visual companion workflow skill

    tianyuegithub/visual-companion-workflow-skill Skill

    0★

    Use when a user asks for 可视化伴随, visual companion, browser whiteboard, live HTML design discussion, opening a local URL, or iterative visual brainstorming before formal docs or implementation.

  • Dumb army

    ninja-automacoes/dumb-army Skill

    0★

    Use when you need to know if an interface or artifact (web page, app screen, dashboard, PDF, email, Figma export, or raw copy) is easy to understand (low cognitive load / "burro-first"), where real users get confused, or before shipping copy/UX changes.

  • Test gen

    ZaneL1u/test-generate-skills/skills/test-gen Skill

    0★ repo

    Generate frontend tests by deciding which parts belong in unit tests and which belong in end-to-end (e2e) tests, then writing runnable test code that matches the project's existing stack. Framework-agnostic: works with Vue or React, Vitest or Jest, Playwright or Cypress, with or without MSW. 生成前端测试:把要测的内容拆开,判定哪些写单测、哪些写 e2e,再按项目现有技术栈产出可直接运行的测试代码。框架无关,自动适配项目已有的 runner 与组件框架。 Use this skill whenever the user wants to: - "/test-gen", "生成测试", "写测试 / 写单测 / 写 e2e", "补测试", "add tests", "write tests for this", "how should I test this" - paste a feature description, requirement, user flow, test case, or source file and ask for tests - ask "should this be a unit test or e2e", "帮我拆一下测试", split testing responsibilities Trigger even when the user does not explicitly say "unit" or "e2e" — any request to add automated verification for frontend behavior should use this skill.

  • Aiops vault template

    LinLin00000000/aiops-vault-template Skill

    0★

    Use this skill whenever the user asks to deploy, operate, troubleshoot, migrate, document, audit, inventory, back up, or clean up local/cloud services, servers, Docker/Compose apps, domains, DNS, reverse proxies, SSH entries, service directories, secret locations, or the AIOps vault itself. This skill treats the local AIOps vault as the source of truth, uses low-token command interfaces before full-file reads, enforces secret and high-risk boundaries, and records verified changes back to the right file layer.

  • Codebase trellis

    Shaelz/codebase-trellis-skill/skills/codebase-trellis Skill

    0★ repo

    Use when staging, committing, pushing, reviewing PRs, checking CI, finishing branches, or auditing Git/GitHub repo posture for AI-assisted work. Inspect first, propose second, mutate only after explicit approval. Good triggers: dirty working tree, commit planning, push planning, PR prep, branch cleanup, worktree setup, Git state review, CI inspection, repo posture audit. Trigger phrases: commit, push, stage, PR, branch, worktree, finish, recover, git status, git audit, trellis.

  • Codex guides

    ojesusmp/codex-guides/plugins/codex-guides/skills/codex-guides Skill

    0★ repo

    Create verified project documentation from the real repository: a self-contained visual GUIDE.html for operators, a cited DEVELOPER-INDEX.md for maintainers, and optionally AI-CONTEXT.md. Use when Codex is asked to document a project, create a user guide, build a developer index or technical reference, generate the two companion guides, or prepare durable context for a future coding session.

  • Issue to agent

    itscloud0/issue-to-agent/.agents/skills/issue-to-agent Skill

    0★ repo

    Turn a GitHub issue plus the current local repository into a scoped implementation task pack for Codex.

  • Agent memory palace

    frankxai/mind-palace-agent-skills/skills/agent-memory-palace Skill

    0★ repo

    Give an AI agent persistent, navigable memory across sessions using a memory-palace structure. Use when building agent memory, when the user says "remember this across sessions", "give the agent long-term memory", "persistent context", "knowledge graph for my agent", or for coding agents, Claude cowork, and ChatGPT projects that must retain state. Reads and writes palace.json as a durable spatial knowledge store the agent reloads each session.

  • Claude md

    WaiYanNyeinNaing/claude-md-skill/skills/claude-md Skill

    0★ repo

    Best practices for writing, structuring, auditing, and improving CLAUDE.md context files for Claude Code projects. Use this whenever the user wants to create a new CLAUDE.md, review or audit an existing one, set up Claude Code configuration for a project or monorepo, trim a bloated CLAUDE.md, or set up progressive disclosure with @imports / .claude/rules/ / .claude/skills/. ALSO use this for agentic and long-running engineering — production AI projects where an LLM calls many tools/APIs, work spanning many Claude Code sessions over weeks or months, persisting checkpoints/rules/session state across sessions, memory layers (CLAUDE.md, MEMORY.md auto-memory, the Memory Tool, subagent memory), surviving context compaction, or the initializer + coding-agent harness pattern — even when the user only says something like "set up Claude Code for this repo," "how do I keep context across sessions," "my agent forgets things between runs," or "why does Claude keep ignoring my instructions."

  • Multi agent project skill

    aduhappy/multi-agent-project-skill Skill

    no license0★

    初始化或重构"跨 agent、跨软件可续接"的项目文档骨架。当用户开新项目/新课题、想建立多智能体协同工作流、提到 AGENTS.md / CLAUDE.md / GEMINI.md / .cursorrules / copilot-instructions / 项目记忆 / agent 接力 / 换软件继续 / 上下文丢失,或现有项目缺统一入口导致每个 agent 反复问相同问题、互相覆盖、接不上前任工作时使用。即使是"帮我建个项目文档""怎么让多个 AI 工具协同""开个新课题要准备什么"这类模糊请求也适用。

  • Autoresearch

    libenxier-beep/codex-custom-skills/skills/autoresearch Skill

    0★ repo

    Use when the user wants Codex to run, adapt, or audit a bounded metric-driven experiment loop for code, CLI behavior, prompts, benchmark harnesses, or karpathy/autoresearch-style repos with clear editable scope, validation command or eval set, metric, and keep/discard criterion.