Engineering
98,276 rows, by stacks then stars
ponponusa/agent-context-maintainer Skill
1★Create, update, and validate repository-local AI agent context files such as AGENTS.md, .agents/core.md, .agents/routing.md, provider-specific profiles, and task SKILL.md routing. Use when an AI coding agent (Codex, Claude Code, Gemini, Copilot, Cursor, or others) is asked to make AGENTS.md or SKILL instructions self-maintaining, inspect the current path and generate optimal agent context for it, or keep agent instructions synchronized with the repository's actual structure.
Mahermenix/vault-sync Skill
1★Use at the end of any agent coding session to reflect what happened back into the Obsidian vault. Reads the session transcript and git history, extracts decisions, tasks, projects, people, research, and other entities, then creates new vault pages or updates existing ones using the right templates and folder placement. Maintains the vault's zero-orphans principle by updating index files and adding wikilinks. Use whenever the user mentions ending a session, wrapping up, syncing to vault, logging the session, updating Obsidian, capturing what we did, saving session notes, or says "/vault-sync", "/end-session", "/log-session" — and after long working sessions even if they don't explicitly ask. Supports planning mode (capture what they intend to do) and work-log mode (capture what actually happened); work-log is the default.
mohamedzhioua/proofguard/deferred/api-contract-guard/skill Skill
1★ repoBlocks a breaking change to a published contract — a REST/RPC/GraphQL endpoint or its request/response fields, an event/webhook payload schema, or a public library signature — made IN PLACE (renamed/removed/retyped field, newly required param, changed status code/error shape, tightened validation) instead of via expand-contract (add the new shape alongside the old, migrate consumers, deprecate with a window, remove later) and/or a version bump. Best used reactively before changing any endpoint, event schema, or exported function signature with existing consumers. Use when the user says "rename this field", "change the response shape", "make this param required", "update the API", "change the payload", "return a different status code", or after an agent edits a contract other code already calls. DO NOT USE for doc sync (docs-drift-guard), diff scope (clean-diff-guard), or the done-claim (evidence-before-done); this guard is specifically about not breaking existing consumers of a published contract.
shinpr/grok-workflows/plugins/claude-to-grok/skills/claude-to-grok Skill
1★ repoMaps Claude Code tool, agent, and progress-tracking identifiers to Grok runtime equivalents while preserving claude-code-workflows (dev-workflows*) execution semantics. Use when running dev-workflows* on Grok, including recipe-* orchestration that needs CC→Grok translation.
3bsalam-1/architecture-extractor Skill
1★Analyze a codebase and generate a highly detailed `docs/engineering_constitution.md` that captures its engineering DNA — architecture philosophy, coding standards, patterns, conventions, and structural rules — in a reusable form that can be applied to a completely different product. Use whenever the user asks for: an engineering constitution, architecture extraction, coding standards document, project engineering handbook, codebase DNA, reusable architecture template, "extract the patterns from this codebase", or "document how this project is built". Also trigger when the user wants to start a new project "in the same style as" an existing one, or wants to capture the implicit conventions of a repo before a major refactor or team handoff. This skill is about HOW the project is built, never WHAT it does.
arlegotin/pythurion/skills/pythurion Skill
no license1★ repoImproves correctness, maintainability, typing, reliability, testability, and measured performance for Python code, optimized for Python 3.12 and newer. Use whenever a task reads, writes, explains, reviews, refactors, debugs, tests, designs, optimizes, profiles, or benchmarks Python—including source files, snippets, code blocks, tests, notebooks, scripts, packaging, project configuration, or workload latency, throughput, CPU, memory, allocation, startup, and I/O—even when quality or performance guidance is not requested. Do not use when Python code, Python runtime behavior, and Python project configuration are all incidental to the task.
Paldom/python-skills/.claude/skills/add-skill Skill
1★ repoAuthors or improves an Agent Skill in this repo via the eval-first workflow - scope one purpose, write trigger evals, draft SKILL.md, validate, iterate. Use when the user asks to create, add, write, refine, or fix a skill ("add a skill for X", "improve triggering of Y"). Not for scaffolding a new repository or editing infrastructure (hooks, CI, the validator/scripts, docs).
The same file is in 11 other repositories
rajshah4/ai-framing-skills/diagnose-use-case Skill
1★ repoUse when the user has an AI system that is already running (production, pilot, or serious testing) and wants to know why it underperforms, whether it's actually working, or whether to keep going, change course, or kill it. Triggers include "why isn't this working", "metrics are flat", "users aren't adopting it", "should we keep investing in this", or any post-launch review. For projects not yet built, use frame-use-case instead.
leksman/ai-security-guard Skill
1★Harden and audit an app's AI/LLM features against prompt injection, jailbreaks, privilege escalation, action-marker/card forgery, PII and cross-tenant leakage, and media/vision abuse. Use when adding, reviewing, or securing any feature that sends user-controlled text, voice transcripts, or images to an LLM. Installs runtime guards (injection detection, context sanitization, marker stripping), wires blocked-attack alerting (Sentry or any sink), and can run a multi-agent security audit against the codebase.
wha7ev9r/rsshub-route-skill/rsshub-route Skill
1★ repoCreate complete, PR-ready RSSHub route code — namespace.ts, route handler (API/HTML/Puppeteer), Radar rules, and PR checklist. Use whenever the user wants to build an RSS feed for a website via RSSHub, add/contribute a route to DIYgod/RSSHub, convert a site or API into an RSS source, or turn website updates into a subscribable feed. Trigger on: "RSSHub", "RSS route", "rsshub route", "RSS feed for X", "为网站制作 RSS", "给 X 做 RSS 订阅", "订阅 X 的更新", namespace.ts, radar rules, route handler, DataItem, or requests to monitor site updates as RSS — even without explicitly saying "RSSHub". Also trigger when the user is already writing RSSHub route code and needs help with cheerio selectors, ofetch, cache.tryGet, parseDate, Puppeteer, or the PR submission checklist. Do NOT use for: deploying/hosting RSSHub instances, debugging existing route runtime errors, tuning cache/config, general RSS reading, standalone non-RSSHub RSS projects, or non-RSSHub scraping.
michaeljabbour/amplifier-skill-forge Skill
1★Spin up, fan out, drive, and test terminal apps in persistent PTY sessions via the forge terminal daemon — including orchestrating coding agents (Claude Code, Codex, Gemini, OpenCode) and Microsoft Amplifier. Use when asked to launch or test an app in a terminal, drive a TUI, run terminal sessions in parallel, operate Claude Code / Codex / OpenCode / amplifier programmatically, spawn or delegate to another coding agent, coordinate multiple agents through shared artifacts, babysit long-running pipelines, or when you need to launch, drive, or test Amplifier itself in a real terminal session (interactive runs, or supervising attractor / evaluation / resolve / superpowers jobs via their file ledgers).
SamsShow/powerclaw Skill
1★Frontier operating discipline for Claude on any model, plus proactive loop and automation suggestions. Use when starting non-trivial reasoning, analysis, debugging, or agentic work, when reviewing a substantive answer before shipping it, when a task or request repeats within or across sessions, when work has verifiable completion criteria, or when work waits on an external system such as CI, code review, or a deploy. Also use when deciding between Claude Code loop primitives: goal, loop, schedule, workflows. Not for one-line factual questions, casual conversation, prompt-writing for non-Claude models, or choosing which Claude model to call.
jpoindexter/ux-flow-skills/skills/flow-ai-chat Skill
1★ repoProven UX flows, user journeys, and app layout patterns packaged as agent skills for Claude Code, Codex, Gemini CLI, and any coding agent
dkmqflx/fastapi-best-practices-plugin/plugins/fastapi-best-practices/skills/fastapi-best-practices Skill
1★ repoUse when writing, reviewing, or refactoring FastAPI code — endpoints, APIRouter, query/path parameters, dependencies (Depends), Pydantic request/response models, error handling, async def vs def, lifespan events, streaming, background tasks, settings, middleware, or CORS. Triggers on FastAPI backend work, route design, or API code review. For auth/security see fastapi-security; for tests see fastapi-testing.
The same file is in 1 other repository
mifochen/quantall-mcp/QuantAll-mcp Skill
no license1★ repoQuantAll(全A解析)MCP —— 股市全市场向量化计算引擎,为 AI 提供本地 Python 计算环境。 AI 编写向量化代码片段,数秒内完成 5000+ 股票的因子计算、策略回测、IC 分析和 GPU 可视化。 让 AI 从"信息查询者"升级为"数据计算者"——用代码算出客观结果,而非搬运网络观点。 触发:用户明确提到"回测""因子分析""IC分析""选股策略""QuantAll""全A解析"等量化关键词时。 不主动在普通股市聊天中触发,仅在用户有明确量化分析需求时使用。 能力声明:本技能需在用户电脑上创建 Python 虚拟环境(300MB+)、安装 quantall 库、 修改 MCP 配置、启动本地 HTTP 服务(localhost:8686)、创建配置文件和启动脚本。 所有涉及用户电脑的操作,AI 必须事先告知用户并获得同意。 UpdateStock 为可选辅助 MCP(数据库管理,需 tushare API),非 QuantAll 必需。
The same file is in 2 other repositories
aetox-skills/create-agent Skill
1★Use this skill to design or modify AI agents with a scope-first process. Clarifies mission, boundaries, authority, inputs, outputs, risks, metrics, tools, tests, and stop conditions before producing an Agent Spec.
BuildContext/fable-orchestrator/skills/orchestration Skill
1★ repoFable Orchestrator · universal Grok volume agent — session on Fable 5 architects; Grok 4.5 types and self-selects research/explore/critic/device phases for any scenario. USE WHEN implementing, exploring, verifying, writing five-part specs, consulting fable-advisor, fan-out parallel lanes, self-routing subagents, or managing token spend under Fable + SuperGrok Heavy.
Elio2000/herdr-peer-review Skill
1★Drive a second ("peer") coding agent from Claude inside a herdr session, in a watchable TUI pane, so the user can SEE the multi-agent conversation. Use when running inside herdr and the user asks to have another agent review the current changes / a diff / a file, or to consult / ask another agent ("让 codex/grok review", "问一下另一个 agent", "找 X 咨询一下", "peer review", "第二意见"). The peer agent is a single swappable config knob — codex today, grok / cursor / claude / etc. tomorrow. Completion is detected via herdr's agent-status (agent-agnostic). The peer is ALWAYS launched auto-approve ("YOLO") so the loop never stalls on a prompt and can run unattended / AFK; the peer can also arbitrate decisions to close the loop with no human present. Encodes the project's Claude = Code Agent / peer = Reviewer split.
cskwork/build-intuitive-rca-wiki Skill
1★Turn technical incidents, rollback reports, root-cause analyses, postmortems, and system failure explanations into intuitive standalone HTML wikis for nondevelopers. Use when a user asks for a visual or easy-to-understand incident explanation, a beginner-friendly RCA page, a reusable one-file HTML wiki, or when raw IDs, table names, error codes, and architecture diagrams need a plain-language presentation layer.
Oak-Ridge-Ventures/openclip-skill/skills/openclip Skill
1★ repoTurn long videos into short, captioned viral clips from your agent via the OpenClip MCP server. Also FREE with just an account (no subscription): transcribe a video, convert/compress/trim/crop/resize/mute a video, extract thumbnails, edit an image, remove an image background. Plus generate a short UGC-style ad clip from a brief. Triggers include "clip this video", "make shorts", "repurpose this", "find viral moments", "transcribe this", "convert to mp4/gif/mp3", "compress this video", "remove the background", "make a UGC ad", "openclip".
holapancho/packaging-datacloud-skill/skills/packaging-datacloud Skill
1★ repoPackage and ship Salesforce Data 360 (Data Cloud) Data Kits as managed 2GP packages: Standard vs DevOps kit types, retrieve workflow, SSOT dependency, publishing sequence, KQ cleanup, deploy after install, undeploy/uninstall (including Currency Connection / StaticCurrencyRates blockers), and packaging oddities. Use when packaging data kits, DataPackageKitDefinition, DMOs, data streams, or Data Cloud metadata for AppExchange or managed package distribution.
zhongmaomao/grill-with-tree/skills/grill-with-tree Skill
1★ repoInterview the user through a visible decision tree, inspect available evidence before asking questions, trace downstream and cross-branch consequences, and produce an implementation-ready decision brief with assignable tasks. Use when a user asks to grill, challenge, stress-test, align, or de-risk a product, architecture, migration, or implementation plan before work begins, especially when they want a persistent decision record or a handoff to a multi-agent workflow.
giacomoguidotto/skills/skills/agentic-os Skill
1★ repoCoordinate versioned cross-System capabilities. Use when the user invokes /agentic-os scout, /agentic-os pursue, or /agentic-os upskill.
The same file is in 2 other repositories
Igapyon miku ai assistant builder
igapyon/miku-ai-assistant-builder-skills/skills/igapyon-miku-ai-assistant-builder Skill
1★ repoベータ版。Microsoft 365 Copilot内の軽量なAgent BuilderまたはGoogle GeminiのGemに投入するデータを準備し、雑多な既存フォルダを二段階で配備用フォルダへ変換するスキル。Agent Builderを主対象とし、Gemにも対応する。第1段階でJSON / JSONLを1入力1XLSX候補、それ以外の同梱`miku-text-bundle`対応テキストをバンドル候補として確定し、人間の資料追加を待つ。第2段階ではJSON / JSONLの明示mappingをレビューして固定し、確定条件から再実行可能なNode.js変換ジョブを作る。Agent Builder向けMarkdownをDOCX化し、GemではMarkdownのまま使うか同梱`miku-md2docx`でDOCX化するかを利用者が選ぶ。構成固定の内容更新は同じジョブを再実行し、前回条件を再検討する反復実行では新しい作業フォルダを作る。ファイル添付を利用できるライセンス、アカウント、管理者設定が前提。`igapyon-miku-ai-assistant-builder`または`miku-ai-assistant-builder`が明示されたとき、このスキルを使ったフォルダ変換、その再開、確定済み構成の内容更新、または前回条件を参照した再実行が依頼されたときに使用する。Copilot Studio固有のエージェント作成には使用しない。
ashllll/manage-xiao-esp32-projects Skill
1★Create, migrate, document, archive, automate, and validate Seeed Studio XIAO ESP32C3/ESP32S3/ESP32C6 projects using PlatformIO, ESP-IDF, MkDocs Material, official Espressif MCP services, and GitHub Actions. Use for scaffolding or maintaining a XIAO ESP32 project, selecting boards, researching peripherals, documenting pins and APIs, configuring Flash/PSRAM, auditing local ESP32 references, removing unreliable derived documentation, generating code from a specification, building firmware, preparing CI/docs, or installing and improving this skill across Codex, Claude Code, Cline, GitHub Copilot, Gemini CLI, Kimi Code CLI, OpenCode, Pi, and Reasonix—especially on Apple Silicon Macs.
Paldom/node-skills/.claude/skills/add-skill Skill
1★ repoAuthors or improves an Agent Skill in this repo via the eval-first workflow - scope one purpose, write trigger evals, draft SKILL.md, validate, iterate. Use when the user asks to create, add, write, refine, or fix a skill ("add a skill for X", "improve triggering of Y"). Not for scaffolding a new repository or editing infrastructure (hooks, CI, the validator/scripts, docs).
The same file is in 11 other repositories
n24q02m/agent-chat-plugin Skill
1★Use when two or more agent sessions (Claude Code or other tools, same machine or peers) must coordinate through markdown files in a shared folder instead of one supervisor driving subagents — group chat, blackboard, mailbox/inbox, multi-session handoff, or an agent-chat/ directory. Triggers include peer agents, N sessions sharing a folder, waiting for another session's reply, atomic task claiming, and needing multiple separate group chats/channels.
piyushomanwar16/vibe-app-security-audit/skills/vibe-app-security-audit Skill
1★ repoSecurity audit and hardening for vibe-coded apps and AI-generated codebases (Lovable, Bolt, Replit, Cursor, v0, Next.js, Supabase, Express, Firebase). Use when preparing an app for launch, reviewing AI-generated code for security, "check my app for vulnerabilities", "secure my app before deploy", "harden this codebase", or after any major feature update. Profiles the app, then runs a five-phase loop (secret leaks, PII flow, pre-deploy hygiene, deep logic audit, attacker simulation) that flags every issue with file:line + severity, fixes it, and reports exactly what changed.
shenyuhao1/agent-doc-generator/skills/agent-doc-generator Skill
1★ repoCodex and Claude Code skill for turning development sessions, conversation transcripts, or project notes into illustrated, reproducible Word/DOCX development documentation. Use when the user wants automated documentation with Playwright screenshots, precise code-rendered diagrams, or agent-generated explanatory visuals, especially when Chinese text and image encoding must remain reliable. Do not use for editing an existing DOCX; use a DOCX editing skill instead.
nathan8823/fairbill/.claude/skills/benchmark-price Skill
1★ repoEstablish the fair price for billed services using Medicare rates, the hospital's own transparency file, and regional benchmarks. Use after review-bill, before making any offer.
HiIamZeref/design-partner/skills/design-partner Skill
1★ repoConducts architecture discovery and system design sessions — the phase before Spec-Driven Development. Structured brainstorming, state-of-the-art research with current costs, multiple approaches with explicit tradeoffs, decision-by-decision alignment, and final consolidation into an ADR-style decision document ready to feed a Constitution and PRDs/specs. Use this whenever the user wants to brainstorm a new system, service, or feature; compare architecture approaches; discuss design tradeoffs; ask "how does the industry solve X"; or consolidate a technical discussion into a decision document — even if they never use the word "architecture". Do NOT use it for one-off technical questions that have a direct answer.
shahrulestar/skills/skills/ab-testing Skill
1★ repoGeneral-purpose A/B testing (split testing) for any project built or maintained by coding agents: websites, web apps, mobile apps, dashboards, and edge/serverless stacks. Covers hypothesis design, sample-size math, stack-aware variant implementation, event tracking, low-traffic alternatives, pre-flight review, honest analysis, and ship-or-no-ship decisions. Use whenever the user wants to test a UI/UX change, compare two versions of a page, feature, button, headline, onboarding flow, pricing display, or copy; run an experiment or split test; do conversion rate optimization; or asks "which version performs better" or "should I change this" — even without saying A/B test. Also covers agent-specific advantages: reading the actual codebase to produce mergeable variants, persona pre-flight before spending traffic, wiring bucketing logic, and writing statistical analysis when results arrive. Works across Next.js, React, Vue, Workers, React Native, and greenfield repos with no prior experiment setup.
xiaopixuan/skills-public/skills/drawio-diagram-skill Skill
1★ repoUse when the user requests diagrams, flowcharts, architecture diagrams, ER diagrams, UML / sequence / class diagrams, SysML / MBSE diagrams (block definition, internal block, requirement, parametric), BPMN business process diagrams, swimlane / cross-functional flowcharts, network topology, cloud architecture from Terraform or Kubernetes manifests, ML/DL model figures (Transformer/CNN/LSTM), mind maps, or any visualization. Also use proactively when explaining systems with 3+ components, complex data flows, or relationships that benefit from visual representation. Best suited when the diagram needs custom styling, rich shape vocabulary, swimlanes, or exportable images (PNG/SVG/PDF/JPG). Generates .drawio XML and exports locally via the native draw.io desktop CLI.
YouRen1320/agent-project-kit/.agents/skills/api-contract Skill
1★ repoReview and plan API contract changes across producers and consumers. Use for endpoint paths, methods, request or response fields, pagination, errors, uploads, status values, authentication context, or frontend/backend contract mismatches such as 404 and undefined fields.
MuKaramia/PaperFlow Skill
1★Legacy compatibility entry for PaperFlow. Route PDF archiving, translation, annotation, and connected literature-note tasks to PaperFlow Atlas; route research-question synthesis, literature-review reasoning, candidate-paper integration, versioned revision, and collaborative academic writing to PaperFlow Thread. Prefer installing the complete PaperFlow multi-Skill plugin.
AmsonntagChow/ratemyharness/plugins/ratemyharness/skills/ratemyharness Skill
1★ repoUse this skill to rate, audit, red-team, or release-gate a concrete AI agent harness, runtime, orchestrator, runner, repository, deployment, trace set, or configuration. Use for loop correctness, tool dispatch and result correlation, context and memory isolation, permissions and sandboxing, approvals, retry and idempotency, timeouts, cancellation, budgets, termination, observability, cost, recovery, author defense, and same-rubric re-reviews. Trigger for wording such as rate my harness, audit this agent runtime, review my agent loop, red-team this orchestrator, is this agent production ready, harness 上线前挑刺, 给 agent harness 打分, or 这个智能体运行时能上线吗. Require an actual harness artifact or runtime evidence, including one in the current workspace. Do not use for generic agent architecture advice, ordinary code review, reviewing a standalone Skill or prompt, validating only a plugin manifest, debugging one model response, or evaluating a model without its runtime.
PolicyLayer/mcp-precheck/skills/mcp-precheck Skill
1★ repoCheck MCP servers against the PolicyLayer registry before connecting to them. Use when you are about to add, install or configure any MCP server (editing .mcp.json or a client config, running `claude mcp add` or an npx-launched server), when your human asks you to evaluate or vet an MCP server, when asked to scan or review the MCP stack already configured on this machine, or when asked to install or remove the precheck hook.
no7z/skillstat Skill
1★Audit which installed Claude Code, Codex, and Cursor skills actually fire, compare usage across agents, estimate Claude context cost, and safely slim unused Claude user skills. Use when the user asks which skills they use, which agent used them, whether they have too many skills, or what is bloating context. Runs the local skillstat CLI offline.
gpt-img-2/gpt-image-2-ecommerce-skill/skills/gpt-image-2-ecommerce Skill
1★ repoPlan, prompt, and review product-preserving ecommerce image sets for GPT Image 2 or other image models. Use when the user needs marketplace main images, white-background product photos, product detail page or PDP sequences, lifestyle scenes, feature graphics, ad crops, product-image prompt packs, or QA against source product references. Also use when generated product images change geometry, color, material, labels, logos, controls, ports, packaging, accessory counts, or verified claims.
dineshrevunuru/SuperSkills/analyze-usability-data Skill
1★ repoTurns raw usability-test data (session notes, recordings, task outcomes) into prioritized, severity-rated findings using NN/g's 4-step analysis flow, a 6-dimension data-quality filter, and Nielsen's 0–4 severity scale (frequency × impact × persistence). This skill is the CANONICAL HOME of the severity scale — run-heuristic-evaluation and audit-accessibility rate their findings through it. Use when a usability test has been run and someone asks to 'analyze the usability data', 'what did we learn from testing', 'write up the findings', 'how bad is this issue', 'rate the severity', or 'prioritize these usability problems'. NOT for synthesizing generative interviews (synthesize-research-data) or statistical claims from large samples (use-quantitative-evidence).
adamhjort/lovable-porting-agent Skill
no license1★Inventory, remediate, and port Lovable-generated applications to independently operated infrastructure. Use for Lovable export or migration, portability assessment, pre-production recreation, optional approval-gated database cloning, removal of Lovable runtime dependencies, or repeatable deployment planning. Select hosting and backend independently across edge, static, serverless, container, Supabase, self-hosted, and common PostgreSQL targets. Works with Agent Skills clients including Codex and Claude Code. Never copy data implicitly, expose secrets, or tear down the source environment.
kayaman/agentskills/.agents/skills/semver Skill
1★ repoEnforces Semantic Versioning 2.0.0 (semver.org) rules. Use when choosing a version bump type for a release, validating version strings, tagging releases, handling pre-release or build metadata, managing deprecations, or advising on version precedence.
dorukardahan/headless-relay Skill
1★Headless handoff guide for running other AI models from inside an agent session (any Agent Skills runtime - Claude Code, Codex, Grok Build, Cursor, OpenClaw, Hermes). Covers GPT (codex exec), GLM (opencode run or zcode --prompt), Grok (grok -p), Gemini (Antigravity agy -p), and Claude (claude -p or a subagent) - inline vs file prompts, parallel multi-model consensus, JSON output, session resume, image/video generation, provider-terms compliance. Use for "ask codex", "ask GLM", "ask grok", "ask gemini", "second opinion", "cross-model review", "generate an image", "run headless", "ask another model".
wlj103/super-prompt Skill
1★AI prompt engineering tool: anti-ambiguity interception × web pre-search × layered guidance × seven-gate pipeline. Turns one sentence into production-grade prompts for Claude, GPT, and AI agents. Supports system prompts, SOPs, agent instructions with multi-format delivery (diagrams, podcasts, docs). Trigger: write prompt, refine prompt, optimize prompt, write system prompt.
ycmjason/aibridge/skills/aibridge Skill
1★ repoDrive models from OTHER providers as your delegates — bridge a task to another AI CLI on this machine: an orchestrator-driven plan → implement → review workflow across Grok / Gemini / Codex / Claude, one-shot cross-model delegation, and capabilities your own provider may lack, like real image generation on a Codex, Antigravity, or Grok seat. Runs on the backing CLIs' existing logins — no API keys. Use for (a) creating / generating / redrawing / restyling an image / icon / graphic / illustration, or writing an image-gen prompt; (b) delegation — and reach for this PROACTIVELY, before implementing a sizeable, well-defined, self-contained chunk yourself: `subagent` hands any clearly-specified task to another model (concurrent, with different training and blind spots than yours — cross-model second opinion / red-team / long-context analysis); (c) sizeable or risky implementation work — `plan` has a model expand your intent into a detailed plan FILE against the real codebase, you read/approve/edit it, `implement` executes it in place running the real gates, and `review` cross-checks the resulting diff against the plan contract (over-reach is a finding) or pre-reviews the plan before any code is written. File PATHS, not contents, travel between stages — cheap on your output tokens. Models use canonical effort-aware slugs (`xai-grok/grok-4.5`, `google-antigravity/gemini-3.6-flash`, `openai-codex/gpt-5.6-sol-high`, …) — no short aliases, always the full slug. Quota is relative to whoever runs this skill: a backend that shares YOUR own quota (`anthropic-claude/*` for Claude-based agents — it bills the claude CLI's subscription — `google-antigravity/*` for Antigravity-based agents, …) is a last resort; say so when you reach for it.
manuelmauro/arkouda/skills/prepare-release Skill
1★ repoPrepare a new arkouda release with changelog and version bumps
chemny/agent-skill-manager Skill
1★Use when the user wants to list, search, inspect, audit, update-check, activate, deactivate, report on, or manage locally installed agent skills and adjacent capabilities across Codex, Claude Code, OpenClaw, Hermes, and configurable local roots.
Additive agent tool surface construction
vishuwa2004/cskill-agents/agents/claude-code/skills/additive-agent-tool-surface-construction Skill
1★ repoMerge agent-specific tool sets (MCP clients + fetched tools) into the parent context without disrupting the parent’s existing pool.
The same file is in 1 other repository
zj021033/senior-mode Skill
1★Make your coding agent operate like a senior engineer instead of an eager junior — and stop it from cosplaying one. Activate for any real engineering work: debugging, planning, implementing, reviewing, refactoring. Five gates that fire at decision points: (A) a user CLAIM → check it before building on it; (B) a PLAN → pressure-test it against the goal and feasibility; (C) a BUILD request → pin the goal before writing code; (D) something BROKEN → find the root cause, never a band-aid; (E) every reply → cut the slop. Plus a hard boundary that blocks manufactured pushback, needless questions, and ceremony on trivial work. Triggers on: "fix this", a bug or failure, a plan to review, a feature to build, "is this the right approach", or whenever the agent is about to agree, patch, guess, or pad.
jesse-black/docgarden/.agents/skills/agents-md-authoring Skill
1★ repoCreate or revise repository `AGENTS.md` files using a progressive-disclosure, dispatcher-style structure. Use when an agent needs to author or edit a top-level or directory-local `AGENTS.md`, convert a long instruction file into task-oriented "Start Here for ..." sections, add a concise "Must Follow" section with Always/Never rules, improve skill-discovery cues, or decide whether guidance belongs in root `AGENTS.md`, deeper docs, local `AGENTS.md`, or a repository-local skill.
BackEngine-ai/backengine-mcp/ai.backengine/backengine-mcp MCP server
no license1★ repoSurface customer & prospect context from Slack, email, transcripts and tickets in any MCP client.
everway-ai/everway-mcp-server/ai.everway/everway-mcp-server MCP server
1★ repoRead and update your Everway trips and itineraries from any MCP-compatible AI assistant.
Lithtrix — Identity, Memory & Trust for AI Agents
lithtrix/lithtrix-mcp/ai.lithtrix/lithtrix-mcp MCP server
no license1★ repoIdentity, memory, trust, and swarm primitives for AI agents. MIRC base always free.
marchward/marchward/ai.marchward/mcp-server MCP server
1★ repoRuntime authority for AI agents: credential mediation, spend cap, approval gates, audit log.
slidecraft-in/presentations-ai-mcp-server/ai.presentations/presentations-ai MCP server
1★ repoPresentations.AI MCP server — create designed slide decks from a topic, text, or document.
keithah/hostex-mcp/ai.smithery/keithah-hostex-mcp MCP server
1★ repoManage your Hostex vacation rentals—properties, reservations, availability, listings, and guest me…
ramadasmr/networkcalc-mcp/ai.smithery/ramadasmr-networkcalc-mcp MCP server
1★ repoLook up DNS information for any domain to troubleshoot issues and gather insights. Get fast, relia…
wgong/sqlite-mcp-server/ai.smithery/wgong-sqlite-mcp-server MCP server
dormant1★ repoExplore, query, and inspect SQLite databases with ease. List tables, preview results, and view det…
VoxellInc/forge-mcp/ai.voxell/forge MCP server
1★ repoMCP server for Forge, Voxell's hosted text-embedding API. Tools: embed and list_models.
icosaedro-git/toolsnap-mcp/app.toolsnap/toolsnap-mcp MCP server
1★ repoDeterministic AI agent microtools, no accounts/API keys. fetch_extract: 98% token cut. 38 tools.