Shape agent capabilities into your workflow.

mcp-searxng is a standalone MCP server that connects one or more existing SearXNG instances to assistants such as Claude or Cursor. It supports paginated web, news, and article searches, time and language filters, suggestions, instance capability discovery, and conversion of readable URLs to Markdown. Multiple interchangeable replicas can fail over or be queried in parallel and merged. The project is not a SearXNG plugin, and it requires at least a configured SearXNG instance URL.

EvoSkill is an open-source framework for discovering and evolving reusable coding-agent skills. It runs agents on benchmark tasks, analyzes failed trajectories, proposes skill and system-prompt variants, and evaluates them on held-out data. Projects can use Claude Code, Codex, OpenCode, OpenHands, or Goose with CSV question-and-answer sets or containerized Harbor tasks. It suits teams with a meaningful evaluation set and requires the selected agent runtime, model-provider credentials, and supporting environment.

dailypaper-skills collects new papers from Hugging Face and arXiv, scores them against your research interests, and produces a daily reading list grouped by priority. Selected papers can be expanded into structured Obsidian notes covering methods, experiments, equations, figures, limitations, and follow-up questions, with indexes maintained automatically. It requires Claude Code, Obsidian, Python, and Poppler; Zotero integration is optional.

AI Factory configures a project for selected coding agents, installs relevant skills, sets up MCP integrations, and adds spec-driven workflows for exploration, planning, implementation, fixes, review, and quality gates. It supports many agent tools, including Claude Code, Codex, and Cursor. A skills-only installation is available, but it omits the CLI, automatic MCP configuration, and interactive setup.

Factory is a software-development platform built around the Droid coding agent, with workflows available through the CLI, web, Slack or Teams, Linear or Jira, and mobile. Its ecosystem includes a VS Code extension, ACP integration for JetBrains and Zed, TypeScript and Python SDKs, and a GitHub Action for pull-request reviews, security scans, and descriptions. The CLI can be installed on macOS, Linux, and Windows.

Task Observer is a meta-skill that runs alongside normal work, capturing corrections, judgment calls, and repeated workflow gaps. It turns those observations into proposed updates for existing skills or drafts for new ones, so a skill library can evolve from actual use rather than remain frozen. The fullest experience is available in Claude environments with filesystem access, where logs and proposals persist between sessions; web and mobile use a handoff-document mode instead.

Loki Mode turns a PRD, issue, API specification, or short brief into a repository with source, tests, configuration, and audit records. Its agents repeat reason-act-reflect-verify cycles, while deterministic gates block empty diffs, failing tests, and inconclusive checks from being reported as complete. Each run can produce evidence that the completion claim can be rechecked.

The pack lets Claude Code or Cursor drive the Arcads API to create short videos, static Meta ads, thumbnails, and multi-step animated campaigns using model-specific prompt libraries. Basic image and video workflows require Python 3.10+ and an Arcads API key, while captioning, stitching, or publishing pipelines may also need ffmpeg, Node.js, or other CLI tools.

guard-skills supplies focused second-pass reviews for production code, tests, documentation, WordPress, and WooCommerce work. Run the relevant guard after an agent changes a repository to catch patterns such as needless abstraction, ineffective tests, documentation drift, or platform-specific mistakes before commit or merge. The package is primarily Markdown and includes no executable scripts, network calls, or credential requirements.

manim_skill provides separate Agent Skills for Manim Community Edition and ManimGL, with tested examples and guidance for scenes, animations, objects, text and formulas, cameras, and 3D rendering. Coding assistants can load the relevant framework knowledge while working on educational or mathematical animation projects. The two Manim variants are incompatible, and typical use also requires Python, FFmpeg, and LaTeX.

Describe the game you want and start with the director skill; it routes the work across gameplay, graphics, UI, audio, debugging, and release checks. The package includes a Vite and TypeScript scaffold, deterministic test hooks, a seeded random generator, and Playwright templates for browser, visual, and bot testing. Core game work runs without paid API keys, while optional external asset generation requires provider credentials.

SkillSpec audits a skill for discovery risk, context load, buried obligations, undeclared dependencies, and missing proof. It can then import the prose instructions into a neighboring skill.spec.yml contract with explicit routes, forbidden actions, checks, tests, and evidence requirements, while leaving an alignment summary after execution. It is a CLI and contract layer rather than a new agent runtime, orchestration platform, or security boundary.

Use SkillPack to combine skills from GitHub, URLs, or local folders with prompts, policies, personas, and optional scheduled jobs. It produces a lightweight archive that teammates can unpack and launch into a local browser interface, with optional Slack or Telegram access. The runtime keeps team data on the host machine but requires Node.js 22.19.0 or newer and an LLM API key.

Install this collection to give compatible coding agents reusable guidance for Kotlin projects. Each skill packages its own instructions, scripts, and resources, and can be added through the Skills CLI, Claude Code or Codex marketplaces, or copied into a project manually. The current catalog focuses mainly on backend development and tooling rather than providing a standalone application.

These skills guide an agent through builds, signing, TestFlight distribution, metadata and localization, submissions, Apple Ads, and related asc CLI workflows. They also document command selection, authentication, pagination, and JSON-first automation. The pack is community maintained, unofficial, and not affiliated with Apple.

Awesome Android Agent Skills packages Android architecture, Jetpack Compose, data-layer, accessibility, testing, and migration practices as installable Agent Skills. Copy the skills directory and shared Agent.md context into an Android project, then compatible assistants such as GitHub Copilot, Claude, Gemini, or Cursor can load the relevant instructions for each task instead of relying on repeated prompts.

This collection gives Claude Code, Codex, Cursor, and other compatible agents focused guidance for building Apple-platform apps. Its 86 self-contained skills cover SwiftUI, concurrency, data, system frameworks, testing, performance, and other iOS engineering topics, with examples aimed at iOS 26+ and Swift 6.3. Users upgrading from older releases should note that some skill paths were renamed.

Install a collection used by the Sentry team for code review and simplification, commits and branch creation, Django access or performance checks, documentation, API endpoint work, and brand-aware writing. It is available as a Claude Code marketplace plugin or a standard skills package for compatible agents, helping coding assistants follow Sentry's engineering and content practices.