omp packages
-
Cross-harness function complexity gate for coding agents.
-
Enhances pi's native /tree with sticky-left view and a detail pane
-
Usage, rate-limit, and spending monitor for pi coding agent: Codex, Claude, Copilot, OpenCode, and OpenRouter daily spend, key budgets, and account credits.
-
Pi extension and skill for Mnemosyne — local-first AI memory
-
MCP (Model Context Protocol) adapter extension for Pi coding agent
-
Pi coding-agent bridge for Artifact Server: receives annotation bundles as follow-up work and closes them through the comment API.
-
A reusable Pi execution progress overlay and milestone tracker
-
pi 插件:spawn_sub 工具。把多步或上下文很重的任务委派给 tmux 里隔离运行的 pi 子 agent,交付物写到 /tmp/pi-sub-<name>/result.md。
-
OpenAI Responses compaction, Fast mode, prompt promotion, and optional apply_patch for Pi
-
Adds priority service-tier aliases for available OpenAI and OpenAI Codex models in Pi.
-
Browser access extension for the Pi coding agent: web search, URL fetching, GitHub repo cloning, PDF extraction, YouTube video understanding, and local video analysis. Supports OpenAI, Brave, Parallel, TinyFish, Search1API, Searchinfinity, Querit, Tavily,
-
Cursor API key provider for pi.dev - run Cursor local agents from pi using your own Cursor API key, stored only on your machine, with an audited egress surface (Cursor endpoints only)
-
Tiny live TTFT and TPS readouts for the Pi coding agent.
-
Pi tool — drive live Hunk diff reviews with ordered bulk operations
-
A git-native memory tool for AI agents. Personal + team/project memory that any harness (Pi, Claude Code, cloud review bots) can use. Built with Effect.
-
Sandboxed bash execution with parameter-level Git approval, forked from pi-simple-permissions.
-
A read-only branch review command for the Pi coding agent
-
Korean-aware algorithmic conversation compactor for pi (fork of pi-vcc) with session_recall history search.
-
Live + replayable session trajectory for pi — dsh-style turn timeline with TTFT/decode split, backed by pi's session JSONL so history is never lost
-
Material black theme for the pi coding agent - near-black UI, green user messages, and a highlighted final answer so you can see where a run ended at a glance.
-
Installed packages manager — web dashboard for browsing, installing, and removing pi packages from your agent. Adds the /packages slash command.
-
LightPanda browser automation for pi coding agent via CDP. Lightweight, fast alternative to browse97 — no Chrome required.
-
GUARDRAILS.md gate for opencode, pi and Claude Code (hooks module): a System One decision model (TypeSafe Jev) judges every bash command before it runs — destructive / credentials / GUARDRAILS.md policy
-
Privacy-aware AgentPet activity reporter for Pi
-
Pi extension package: fold tool calls and previous replies, localize the native Pi UI, and send a system notification when the agent settles.
-
Reports Oh My Pi token usage (metadata only) to your friend group's tokenmunchers leaderboard
-
Link proposed OMP plans into the current project before review.
-
Workspace-aware simplify, test curation, review, and GitHub pull request workflow for Pi.
-
Bidirectional bridge between pi main conversations and Feishu (Lark): sends a Feishu notification when a pi task finishes, and lets you reply to the notification in Feishu to remotely command that pi session to continue. No lark-cli dependency, built on t
-
Give DeepSeek (text-only) models vision in Claude Code, Codex, and Agent Plugins clients: describe images via any OpenAI-compatible vision endpoint.
-
Typed Power BI Report Server REST API v2.0 client and read-only pi extension
-
Pi extension that manages third-party providers and relay endpoints through models.json
-
Agent teams for pi, ported from claude-code-teams-mcp
-
Hatz AI provider installer for coding agents (omp, pi, hermes, claude-code, openclaw)
-
Run persistent actor-style subagent swarms inside pi.
-
Pi extension that brings Fallow codebase intelligence into Pi as an agent tool and slash command.
-
Pi extension — intercepts destructive agent operations (file delete, Docker, process/service stop, Git, database, orchestrators, install/uninstall) and requires explicit user approval before execution.
-
Batch web search for the pi coding agent -- ranked, readable results ready to fetch. Requires pi-smart-fetch.
-
No More Agents Dot MD: turn .pi/context/*.md files into declarative behavior rules (conditional context injection, tool guards, confirmations, input transforms, notifications) managed with the /nma command.
-
Pi extension — @ file picker overlay with fuzzy + git recency ranking
-
A guided Git conflict-resolution command for the Pi coding agent
-
Declarative natural-language sentinel rules for the pi coding agent: side-loop LLM audits on tool calls, tool results, turn/agent boundaries and context thresholds, with blocking gates and background finding injection.
-
pi extension: vantis.
-
AI inference for your coding agents. Route models, control spending, and fund your workspace.
-
Live elapsed time and a minimal completion timestamp for Pi's terminal UI.
-
Jev-backed tool guardrail for pi
-
Image generation tool for the pi coding agent (Codex / Grok / OpenAI Images)
-
Pi extension that scopes the providers visible in /model based on the working directory.
-
Reviewable implementation plans for the pi coding agent: the comprehensive-planning skill, the standards it routes to, and an extension that triggers, validates, and tracks the plan.
-
Coding profiles: inject a prompt file as a persistent system-prompt mode