Packages: Y
Every listed package whose name, or scope, starts with Y.
- @y0usaf/pi-jev untested TypeSafe Jev as a decision layer for Pi: judges mutating tool calls before they run, reads tool output for secrets and failure class after, and exposes jev_ask for typed calibrated answers instead of prose
- @yaag/extension untested The yaag pi extension: run Orchestration Programs from a pi session.
- @yalieny/pi-better-cost-display-footer untested pi extension: scriptable peak/off-peak tier rules per provider with an enhanced footer (tier label, cache-hit-rate precision, custom currency symbol)
- yamlet-pi untested Spec-driven development with yamlet for the pi coding agent: the yamlet CLI registered as first-class pi tools, with a gate that blocks hand-editing a .yamlet.yaml, .techspec.yaml or .adr.yaml; an interviewing author skill, verifier and refresh skills (Gh
- @yandy0725/pi-ask-user untested Interactive ask_user tool for pi-coding-agent with searchable split-pane selection UI, multi-select, and freeform input
- @yandy0725/pi-coding-tools untested pi package providing AST/LSP code-intelligence tools
- @yandy0725/pi-lark untested Pi package with Lark (Feishu) API skills for AI-powered Lark app development
- @yandy0725/pi-memory untested File-system driven persistent memory layer for pi coding agent
- @yandy0725/pi-sandbox untested pi coding-agent extension: process-level sandbox (bwrap / landlock / seatbelt / windows-acl) — workspace writable, everything else readable, fail-closed
- @yandy0725/pi-subagents untested Focused, in-process sub-agent core for pi — autonomous agents plus a typed API and lifecycle events
- @yandy0725/pi-superpowers untested Superpowers skills for pi — structured development workflows
- @yandy0725/pi-todo untested pi package providing a minimal todo tool with an editor-overhead widget
- @yandy0725/pi-vision-tools untested pi package adding a describe_image tool that lets non-multimodal models delegate image analysis to a vision model
- @yandy0725/pi-web-tools untested pi package providing websearch and webfetch tools
- @yankewei/pi-web-search untested pi extension that exposes provider-native web search for DeepSeek Responses and OpenAI Codex GPT-5.6 models
- @yansircc/pi-chrome untested Let Pi use your existing signed-in Chrome profile through Agent-first tools.
- @yaosu/pi-path-guard untested Path Guard for pi — blocks destructive commands & path overwrites, protects .env/keys, with strict/normal/loose/trusted/naked guard modes (/guard). pi 防误删/防误覆盖扩展,支持 5 种防护模式。
- @yassimba/pi-code-intelligence untested Route Pi code exploration toward Codebase Memory without blocking native tools
- @yassimba/pi-guardrails untested Security guardrails for Pi: protected paths, workspace boundaries, and dangerous-command gates.
- @yassimba/pi-loom untested Loom branding and a startup notice for published Loom updates in Pi.
- @yassimba/pi-loom-mermaid untested Loom's advanced Mermaid renderer for Pi, with better routing and color.
- @yasuhito/clearhead untested A Pi extension that lets the LLM edit its own context to keep the next prompt lean.
- @yasuhito/pi-usage untested A Pi extension that displays Codex and Claude subscription usage plus OpenRouter account credit balance
- @yceachan/pi-better-btw untested pi extension: /btw side-chat overlay forked from pi-side-chat, with main-context sharing and a side-lane guard
- @yceachan/pi-vision-helper untested pi extension + skill: visual understanding via a configurable vision model (pi-registry reuse or custom responses API; default gpt-5.6-luna) when the main model cannot see images — pure TypeScript, single runtime
- @yedan1122/brb untested BRB connects Pi Agent workflows to a real ChatGPT Web conversation as a safe second reasoning workspace. Hand a task out, take the plan back, execute locally.
- @yefengr/pi-reach untested Pair your phone or any browser with the Pi coding agent on your computer. Watch replies as they stream in, send prompts and files, and switch between running Pi sessions.
- @yeliu84/pi-model-router untested Smart per-turn model router for Pi that optimizes your AI budget and usage limits by dynamically switching LLM tiers based on task complexity, budget, and phase awareness.
- @yeyudeh/pi-compaction-files untested Caps Pi's compaction file lists (<read-files>/<modified-files>): drops dead, duplicate and temp paths, hides sensitive ones, strips blocks the summarizer re-seeds into the narrative, and keeps the cost per turn constant.
- @ygncode/pi-web untested Remote control your pi coding agent from any browser on your network
- @yibie/pi-jev-browser untested Isolated Playwright browser for pi, driven by Jev typed decisions through the TypeSafe API or the model pi already has configured.
- @yiki21/pi-agent-team untested Turn several machines running the Pi coding agent into one team over your tailnet. Three transports (broker / mesh / SWIM), one shared API and conformance suite.
- @yiki21/pi-openai-server-compaction untested Pi extension for Codex-style Responses remote compaction, including OpenAI-compatible gateways. Fork of algal/pi-openai-server-compaction.
- @yishan-io/pi-dev-flow untested Pi package containing Yishan's development workflow skills: ideation, planning, debugging, testing, review, subagent orchestration, and whiteboard (Excalidraw scene) editing.
- @yishan-io/pi-subagents untested Pi sub-agent orchestration package for Yishan.
- @yishan-io/pi-task untested Pi tools and workflow skills for SQLite-backed Local Tasks.
- @yoda.digital/pi-remendra untested Persistent memory for Pi — remembers what matters, forgets what's wrong.
- @yorch/pi-claude-delegate untested [DEPRECATED] Delegate work to Claude Code — use pi-harness-delegate (pi install npm:pi-harness-delegate) — shim re-exporting claude harness.
- @youdotcom-oss/pi untested Pi extension for You.com MCP tools and skills
- @youngjurry/pi-agents untested Persistent in-process Codex-style multi-agent collaboration for Pi
- @youweichen/omp-web-ui built for omp OMP Web and desktop interface for Oh My Pi, with a bundled Bun runtime
- @youweichen/pi-harness untested Web chat interface for the pi coding agent, powered by the pi SDK (@earendil-works/pi-coding-agent) — one-command run, Docker/systemd/launchd deployable
- @youweichen/pi-web-ui untested Web chat interface for the pi coding agent, powered by the pi SDK (@earendil-works/pi-coding-agent) — one-command run, Docker/systemd/launchd deployable
- ypi untested ypi — a recursive coding agent. Pi that can call itself via rlm_query.
- @ysskrishna/ai-agent-skills untested Curated Agent Skills (SKILL.md): structured thinking lenses, decision support, and brainstorming workflows for AI coding agents.
- ytdb-slate untested Multi-agent orchestration extension for the pi coding agent: bounded worker sessions, structured episode summaries, and a risk-based review workflow.
- @yuechou/omp-agent built for omp Oh My Pi multi-agent management extension
- @yuechou/omp-display built for omp Oh My Pi extension: sidebar display with session info, files, model, tasks, tools
- @yuechou/omp-memory built for omp Oh My Pi extension: persistent memory, wiki, diary, triples, and graph
- @yuechou/omp-mode built for omp Oh My Pi mode, skill, and extension management system
- @yuechou/omp-video-understanding built for omp Oh My Pi extension: video content analysis using MIMO 2.5 API
- @yugimob/pi-opencode-go-rotation untested Rotate between multiple OpenCode Go API keys with reactive 429, rate-limit, and watchdog stall detection, plus live footer usage for the active key
- @yukikisaku/pi-chain-skills-prompts untested Chain Pi prompt templates and skills into a single prompt
- @yukikisaku/pi-disable-debug-shortcut untested Disable Pi's Ctrl+Shift+D debug shortcut in TUI sessions.
- @yukikisaku/pi-herdr-status untested Herdr tab spinner, completion mark, and AI-generated tab labels for the Pi coding agent
- @yukikisaku/pi-hide-thinking untested Hide thinking blocks and the hidden-thinking label in Pi's TUI.
- @yukikisaku/pi-hide-tools untested Hide tool rows by default and toggle a compact one-line view.
- @yukikisaku/pi-orca-status untested Orca completion marks and AI-generated tab titles for the Pi coding agent
- @yukikisaku/pi-organize-footer untested Organize Pi's footer and toggle extension status items.
- @yukikisaku/pi-plan-build-mode untested Plan/build mode and named preset extensions for Pi
- @yukikisaku/pi-prune untested Prune tool data and thinking blocks from the current Pi session.
- @yukikisaku/pi-retain-cache untested Keep supported Pi model prompt caches warm while the TUI is idle
- @yukikisaku/pi-show-command-list-above untested Render Pi's slash-command autocomplete list above the editor.
- @yukikisaku/pi-ssh-paste-image untested Paste local clipboard text or images into Pi running over SSH.
- @yukikisaku/pi-tmux-status untested tmux window spinner, completion bell mark, and AI-generated session titles for the Pi coding agent
- @yukikisaku/pi-zed-status untested Zed terminal tab spinner, completion bell, and AI-generated session titles for the Pi coding agent
- @yunazgr/pi-companion untested Local-first remote control for Pi sessions: live activity feed, steering, git diff, file drop and phone pairing from a single Rust daemon.
- @yuriteixeira/pi-muxr untested tmux dashboard for Pi coding sessions
- @yuru7/pi-ai-approval untested A fail-closed approval gate for Pi that classifies every covered tool call into six risk levels with an isolated AI reviewer and enforces per-level allow/ask/deny from local configuration.
- @yuru7/pi-metrics untested A Pi extension that prints one dim metrics line when a run settles (elapsed, TPS, TTFT, tokens) and shows the live elapsed time in the working indicator
- @yuru7/pi-native-notify untested A Pi extension that sends a native OS notification when a long-running task completes or when Pi waits for your input
- @yuru7/pi-sandbox untested Process-level, path-transparent sandbox for Pi coding agent with fail-closed workspace protection (bwrap, landlock, seatbelt, windows-acl) and denial-first AI auto-review escalation
- @yusukeshib/pi-advisor untested Explicit, one-shot second opinions for Pi with private reviewable logs
- @yusukeshib/pi-babysit untested Run any shell command and pi subagents under babysit, with context-safe captured output.
- @yusukeshib/pi-colored-model-status untested Color-code the active model in pi's footer so you can tell at a glance which model family (opus, sonnet, fable, gpt, gemini, …) is in use.
- @yusukeshib/pi-stash untested A personal stash of reusable prompt fragments for the Pi coding agent. Push prompts you think of mid-task, then pop them into the editor when you need them.