Directory
Browse agent skills
Search by intent: design systems, testing, deployment, security, documentation, Azure, Supabase, React, or any workflow you want an agent to perform better.
Showing 14,491–14,520 of 126,802 skills
Page 484 of 4,227 · loading 30 at a time
koishi-plugin-sentry-webhook-pusher is a ai & automation skill indexed by Skillful.sh. Package: koishi-plugin-sentry-webhook-pusher.
Probe OpenAI-compatible gateways (LiteLLM, One API, New API, Ollama, ...) for models and metadata
A local kanban board for working in tandem with Claude Code: the kb CLI plus a web board, with an optional GitHub mirror (issues + Projects v2)
Deployment-configured prompt prefix for DeepSeek Harness: ordered system-prompt sections ahead of the persona and reference user/assistant exchanges injected into every ordinary conversation request as real alternating messages (request-path interception,
n8n community nodes for EcoHash: Chat Model for AI Agents, Reranker (BGE) and Embeddings powered by the EcoHash API
WCAG 1.4.12 Text Spacing audit that injects the SC spacing overrides and detects clipped or overlapping text.
WCAG 1.4.10 Reflow audit that narrows the viewport to 320px and measures whether the page requires two-dimensional scrolling.
Inline model capability controls for stock DeepSeek Harness Models settings
Hybrid model router for DeepSeek Harness: declarative rules, user-configurable cost control (cost-first / quality-first / balanced), optional LLM task classifier, and failure fallback chains.
DSH plugin: append each answer's token usage to the conversation
Zro provider extension for pi - GLM-5.2, GLM-5.3 Flash, Kimi K3 & DeepSeek V4 Flash through the Zro inference endpoint
DSH web plugin: desktop notification, chime, and title flash when a session stops — finished output or waiting for human input. Click a notification to jump to that session.
Vendor-neutral control plane for AI-assisted development and AI application lifecycle — one harness, every coding agent (Claude Code, Codex, ZCode, Kimi, Deep Code, Cursor).
Playwright helpers for driving Cheetah Grid in browser automation.
Portable self-improving skill lifecycle with an OpenCode-native review loop.
需求管理 Agent Skill:需求下发 → 分析拆解 → 进度自动统计 → 图文看板汇报,可选 Teambition 双向同步。含标准 SKILL.md(跨 AI 工具可用)与 reqm CLI(终端命令)。
A small, dependency-free REST client for TypeScript — typed responses, pluggable auth, retries, and timeouts.
Persistent background scheduled agent jobs for dsh: cron_add/list/remove/run/logs with durable storage, agent-per-run execution, and a 30s dispatcher tick (Hermes cron + Kimi built-in cron pattern).
Ultra mode for DeepSeek Harness: a per-session effort tier that pins reasoning to max and activates an orchestration policy section, with a composer ULTRA chip and rainbow border in the web UI
Self-hosted plugin repository for DeepSeek Harness (DSH). Each device serves its own repo over HTTP (list/download/upload .tgz, detect non-official plugins), cross-device pull/push, version-aware confirm install, a tabbed '插件仓库' settings page, and repo_*
An integration package connecting WaveSpeed AI and LangChain
Oasis CLI — durable shared local coding-agent sessions with native resume, lazy wake, presence, and in-terminal controls
n8n community node for WaveSpeed AI - generate images and videos with hundreds of hosted AI models
Stop re-pasting your entire repo. Local AI project memory + token cache for Claude, Grok, Cursor, Copilot & more. One brief → every session. Free. Private. Apache-2.0.
TypeScript Result-based API client for Google Search Console API. Type-safe, Valibot-validated, and modern.
Privacy-preserving middleware for embedding pipelines and vector search.
Native OpenCode V2 plugin that gives text-only models vision: relays image parts to Gemini or any OpenAI-compatible provider (OpenAI, Groq, OpenRouter, Ollama...) and replaces them with detailed text analyses before the model dispatch — works with any mod
Hand-drawn editorial poster skill, prompts, and examples · 抽象线条海报生成 Skill