Halofy lifecycle installer and runtime for supported agents; runtime requests are signed with a per-installation Ed25519 key
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 12,271–12,300 of 126,302 skills
Page 410 of 4,211 · loading 30 at a time
Neural Ledger · 神经账本 — turn your DSH collaboration sessions into a visual work ledger: token analytics, smart insights, trend forecasting, and one-click daily/weekly/monthly Markdown reports.
Hash text, files and JSON in n8n with algorithms the built-in Crypto node does not offer, including SHA1, BLAKE2, RIPEMD-160, SM3 and CRC32, plus base64url output and optional salt and pepper.
YuanCe (元测) — a disciplined, authorization-first AI security-testing methodology: a four-stage (recon → discovery → verification → reporting) web security testing workflow on authorized targets (self-owned assets / SRC bug bounty / CTF / local labs), with
An agent organization for Claude Code, structured as a company — 15+ departments, 125+ skills, each independently installable.
Pi extension that registers CodeBuddy (IOA) as an OpenAI-compatible HTTP provider with OAuth/API-key auth
n8n node for sending messages to Microsoft Teams via Teams Bot API
Copilot 引导层插件:WebUI 设置页一键发起 GitHub Copilot 授权(复用 dsh-llm-pi-ai 内置 device-flow),并自动写入 provider profile 激活模型路由。 | Copilot onboarding plugin: one-click GitHub authorization from the WebUI settings card (reusing dsh-llm-pi-ai's builtin device flow) plus autom
Immutable inference contexts, recoverable runs, and warm commitments
生产环境的 skill 自进化系统 | A production self-evolving system for Claude Code: every installed skill keeps accumulating evidence from real usage and evolves when the gate opens - governed, user-authorized, reversible.
Bounded hardware substrate for inference, eval, and checkpointed training. Placeholder — name reserved by Acyclic Labs.
Antigravity-style Teamwork multi-agent orchestration for OpenCode — 10 agents, 6 patterns, 7 slash commands. Replica of Google Antigravity's /teamwork-preview.
Forces periodic Engram memory recall and save reminders for DeepSeek Harness agents: injects a dynamic system-prompt context that reminds the model to load prior session memory at the start of a session, and to save new decisions/bugfixes/learnings every
Framework-agnostic core for the Nubisco CMS: the content contract, field resolution, and the link, image and richtext models. Framework bindings (cms-vue, and later cms-react) build on this.
Evidence-driven multi-agent routing with deterministic safety gates.
Provider-neutral driver, instance, adapter, session, output, and conformance contracts for coding-agent runtimes.
Read your agent CLI session logs, score sessions for cost and friction, and turn expensive patterns into reusable skills, sub-agents, hooks, and instruction files.
This package provides a wrapper for OCI Generative AI services.
Application layer: marketplace, chat, payments, compute, orchestration, and agent tools for BrowserMesh
Docker Micro-Dashboard, AI Container Inspector & 1-Click Prune Plugin for DeepSeek Harness
Portable annotation and interaction contracts for Tileflow maps.
GitHub data aggregation and smart evaluation scoring engine
Turn a repository into a slide deck that tells you when it stops matching the code
Launch anti-detect browser profiles in the cloud and drive them with Puppeteer/Playwright over CDP.
YuanXin (元信) — the pre-install security verifier for Agent skills: deterministic static scan (prompt injection + malicious patterns + SKILL.md integrity + permissions) that outputs a verdict and audited badges. Triggers before installing/evaluating any sk
An anti-AI-slop writing standard and checker for AI-drafted text. Scores any document 0-100 against 43 rules on four axes: plain words, human voice, structure, information density. One JSON rule pack compiles to 18 targets - Claude Code skill, AGENTS.md, Cursor rule, system prompt, Vale style. Zero dependencies, runs offline.
Kubernetes tools for DeepSeek Harness: cluster, namespace, workload, pod, log, rollout, and manifest operations
HTTP client with caching, deduplication, retries and pluggable cache stores (memory, Redis)
smolt - a coding agent that keeps what it learns: curated memory, self-authored skills, and session recall.