/usage for pi: how much of each coding subscription is spent — Claude Code, Codex, Cursor, Copilot, OpenCode Go, Z.ai, OpenRouter
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 14,581–14,610 of 126,802 skills
Page 487 of 4,227 · loading 30 at a time
Model Maker CLI — easy local SFT fine-tuning, quantize, and run for LLMs
React component to easily embed the No-AI scraping protection badge
A zero-dependency JSON parser for Applied-AI-development - LLM streams, featuring VLM coordinate normalization and tool-call standardization.
houserules plugin: a runner-agnostic testing rule, split into opt-in per-language guides.
houserules plugin: changesets integration (canonical changelog) and the optional per-commit changelog ledger.
Run ai-translate against any AI SDK provider — OpenAI, Anthropic, Google, Bedrock, or a gateway — with one interchangeable translation and semantic-audit provider.
File-change monitor & code editor for the DSH Web UI: watch agent edits land as line-by-line diffs in real time, right beside your conversation, then view and edit any workspace file without leaving the browser.
Trace LLM calls with a decorator — non-blocking, fail-silent observability SDK.
n8n node for Klaviyo: profiles, lists, segments, events, metrics and campaigns, plus a trigger for Klaviyo webhooks
Open-source video response SDK for personalized AI applications.
TypeScript SDK for LocalVectorDB — a document-first vector database with SQLite + FAISS, hybrid search, and an HTTP server
Feature-implementation algorithm for ATDD/DDD projects built on the nextjs-fastify stack architecture.
Cloudflare helpers for secure Ooops CMS previews and signed static-site rebuilds.
Add persona thinking, isolated persona querying, and multi-persona debate panels to your agent. For any decision, not just code.
Type-safe HTTP client for Carno.js with automatic codegen
Drop-in flow-board (kanban) view of the content pipeline for content-management-suite-client consumers — composes flow-board-ui with content-aware mapping, headless hooks + default components
TypeScript-first image optimization pipeline for browser uploads.
The runtime under every Fancy connector: modes and estates, connections, one call path, a failure taxonomy that separates "it never arrived" from "nobody can tell", deterministic fakers, webhook verification, pure versioned rendering, declared metric shap
Mazzy Command Center — a full agent orchestrator and command center for the Pi coding agent: durable task tracker, attested orchestration, and a spec<->code<->backlog graph today; its own sub-agent engine, sub-agent creator, meta-agents, and tiered memory
Home Assistant integration for PI — typed tools, code-enforced permissions, no shell access.
Deterministic, non-LLM session titles from a first message — a tiny library plus a ready-to-use opencode plugin.
Native vision for text-only pi models (DeepSeek, Llama, ...). Capability-aware chat attachment interception + vision-describe/vision-ask tools via any OpenAI-compatible vision endpoint (ModelScope, SiliconFlow, OpenRouter, GLM, Ollama), with intent detect
Search arXiv for ML/DL/RL papers and check claims against real abstracts.
Memo — session memory search for DeepSeek Harness agents: cross-session full-text recall (memo_search), distilled notes (memo_remember), and corpus stats (memo_stats) over the official sessionQuery service. Zero infrastructure, local-first.
Live tokens/sec throughput display for the Pi coding agent
n8n community node for enriching company profiles and firmographic data from website domains with the Elvesora Enrichment API.
n8n community node for RankParse - backlinks, domain authority, tech stack, page SEO, and site health from Common Crawl data
Fabric-owned system prompt override and native-prose minimization for DeepSeek Harness