omp packages
-
Delegate background tasks or persistent interactive workers to isolated Pi processes, with execution profiles, bounded delegation, and process-tree cleanup.
-
A compact single-line status bar for pi
-
GPT fast mode for Pi with explicit model allowlists.
-
Codex account rotation extension for pi
-
Interactive ask_user tool for pi-coding-agent with wrapped selection UI, batch clarifications, and freeform input
-
Pi extension for monitoring Z.ai, DeepSeek, OpenAI Codex, and Opencode Go usage
-
Interactive file path autocomplete for pi coding agent - arrow-key navigation, fuzzy matching, Tab completion, and the /pick command
-
Unified web search + content extraction extension for pi with 19 backends (DuckDuckGo, Jina AI, Tavily, Brave, Brave LLM, Exa, Exa MCP, OpenAI Codex, Serper, Firecrawl [keyless], fastCRW, Marginalia, LangSearch, Linkup, Perplexity Sonar, SearXNG, WebSearc
-
A pi extension that replicates GitHub Copilot's manage_todo_list tool
-
OpenAI-compatible provider extension for the Pi coding agent — connects to Rozalia AI
-
Reusable Pi prompt templates for commits, pull requests, and PR review workflows.
-
Poke for pi: detects long-running tool calls and wakes the model when a context compaction stalls the work turn (common with local models).
-
Asana Work Graph tool for Pi. Adds 19 LLM-callable tools (asana_*) that query the Asana REST API over plain HTTP. Mirrors a curated subset of the official Asana MCP tool set — no MCP server install required. Personal access token auth via ASANA_ACCE
-
Multi-account OAuth switching and Claude subscription billing for pi
-
Permission system extension for pi coding agent
-
OCR image-to-text tool for Pi — extracts text from screenshots, terminal output, and code images using Tesseract + ImageMagick
-
In-process tool guard and subagent posture hardening for Pi coding agent. Enforces filesystem allow/deny lists and domain allowlists on read, grep, find, ls, write, edit, fetch_content, web_search, and get_search_content — with interactive ask-tier prompt
-
Metodología ultra como paquete versionado: skills canónicos, registro ADR A/T/P, extensiones de pi y CLI (install/update/status/ship).
-
Pi Postbox extension and web service for remote human decision handoffs.
-
Inline extension commands and composable skill/prompt context with cursor-restoring undo and redo for pi
-
Kimetsu brain as a Pi.dev package — local-first, cross-session memory for the Pi coding agent that gets sharper every run.
-
Opinionated CGx's pi-emote — Animated pixel-art emote widget for pi TUI
-
Custom modes for pi: debugger (read-only RCA) and more
-
OpenCode-style skill loading for Pi: a skill tool, a binding protocol, and archiving of skill bodies that are no longer relevant, with no configuration.
-
Pi extension for a Bifrost LLM gateway: chain models from VK routing rules and/or the full upstream model catalogue, with optional per-family native protocol surfaces
-
Pennix workflow integration of Context Mode: bounded execution, FTS5 search and local session continuity for Codex.
-
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
-
Neuralwatt provider for pi. Shows the reported request cost in pi's footer and provider-reported session energy in the status area.
-
Essential extensions for pi — screenshots, image context pruning, a daily log tool, and a markdown viewer.
-
Pi extension for OpenAI priority service-tier requests
-
Pi provider package for OrcaRouter (GLM 5.3 Flash Free, DeepSeek, Kimi, and more) — OpenAI-compatible gateway with dynamic model discovery.
-
Pi package with a Taiga UI docs skill backed by a self-contained llms-full.txt helper script.
-
Interactive question deck for pi — grill the user round by round: all questions of a frontier in one TUI, recommendation fast-accept, defer, structured answers back to the model.
-
Make the pi coding agent pass Anthropic-compatible gateways that only allow Claude Code clients
-
Your agent says done while your app is on fire. thisisfine locks every behaviour you said yes to into a proven check (browser, API or CLI), and won't let the agent finish a turn that breaks one.
-
Pi tools for reading and safely changing LaunchDarkly feature flags via the ldcli CLI.
-
Always-on subscription usage for Pi: Claude, Codex, DeepSeek, Kimi, MiniMax, Moonshot, Z.AI, OpenRouter and more, live in your terminal as you work.
-
Pi extension for saving and recalling messages across sessions.
-
Native `gd` pi provider for api.girard-davila.net: /login gd accepts a pasted API key or browser SSO, key auto-stored, models appear as gd/<model-id> with no restart. Self-contained (no pi-provider-litellm).
-
Compact Pi extension for recording turnlog provenance with low prompt overhead.
-
Multica workspace health, one command
-
Slash commands that switch models in Pi: bind /quick, /heavy, or your own names to models
-
Instant web search and page fetching for Pi via Monid/TinyFish. No browser, no LLM summarization — just fast structured results.
-
QQ integration for pi — control pi from QQ | pi QQ 集成 — 在 QQ 中操控 pi
-
Folder-based persistent command history for pi. Recall previous commands with shift+up/down across sessions.
-
Pi package with Lark (Feishu) API skills for AI-powered Lark app development
-
Unattended loop mode for pi — separate goal/prepare/run phases with per-phase model selection, endless by default, objective goal function (--check) with score/regression tracking, error retry with backoff, layered anti-repetition defenses (near-duplicate
-
Pi extension that launches a leaderless swarm of pi agents that self-organise on a shared wall, with a live web board.
-
A resilient, adaptive orchestration runtime for Pi
-
Cost report extension for pi