Untested on omp

Context optimization tools for pi coding agent

Pi coding-agent extension — exposes Memongo's durable, cross-project, MongoDB-backed semantic memory as additive tools alongside pi-hermes-memory.

One /fast toggle for Pi: OpenAI priority service tier and Anthropic Claude fast mode (speed: fast).

Minimalist pi extension: live tokens-per-second stats for the active model

Flow-state transition extension for Pi coding agent.

sling extension for delegating tasks to subagents with chains, parallel execution, and TUI clarification

Token, context, and cost usage dashboard for the pi coding agent — a themed CLI plus an interactive /usage command

Improve performance, resource use, determinism, or simplicity through measurement, profiling, benchmarks, and proven equivalence.

Read-only browser inspector for Pi session context and live tool calls.

Interactive Playwright web browsing for Pi. Chromium and Firefox built in; author stealth backends on either engine. Snapshots and screenshots cache to disk, saving context. Configurable profiles/cookies. Agents write site guides, auto-matched by domain.

Automatically update installed Pi packages and Pi itself when a Pi session starts.

IMAP/SMTP email client extension for pi coding agent. Read, search, send, move, and delete emails from your inbox.

PipeLLM provider for Pi with Anthropic Messages, macOS Keychain login, and a management CLI.

A provenance-aware workspace checkpointer for Pi Coding Agent.

Author self-contained implementation plans, multi-plan programs, or compact handoffs that a fresh agent can execute without conversational memory.

Vision capabilities for Pi coding agent — image understanding and visual analysis

Redact sensitive text from Pi provider requests before it leaves your machine.

Pi Plan Mode with optional automatic model switching at Plan boundaries (entry / exit).

Expand %snippet markers in Pi prompts. Comes with built-in fuzzy-finder and trivial tweakability

System One decision models (TypeSafe Jev) for Pi — classify tool for typed probabilistic answers, a Jev-gated permission auto-approve hook for shell commands (default on, static risky list first, never auto-denies, fails safe, observe mode), and a /classi

Read-only REST API client for Pi. Your agent authors and calls YAML recipes handling the tedious parts: pagination, OAuth2, rate limits, backoff. Credentials live in a local store, injected in code, never in agent context. Churn-free research + retrieval

Rewrite rough, plain-language prompts into precise technical prompts before you send them to the Pi coding agent.

Discover, switch, and resume Pi sessions across terminal emulators.

Pi extension that compiles LLM codemode scripts into pi's QuickJS-sandbox JavaScript: adds missing await, strips markdown fences, converts JSON tool-call programs, and translates OpenCode, Codex, Cloudflare, TanStack, Vercel, DeepSeek PTC dialects

Zero-daemon, pure-passthrough native Pi extension connecting directly to OpenCode Zen API

MCP server and Claude Code plugin. Runs an agent's code on fresh, isolated VMs, several at once, each run recorded with evidence

System prompt injection extension for Pi — reads config and appends to system prompt

Fuzzy prompt history for Pi

Auto-classifier Authorizer chain link for pi-permission-system: a light model reviews each ask and returns allow, deny, or defer

Pi extension package for user, project, and package-level tool permission hooks

Gives models awareness of mid-session model switches in Pi

Workflow + project management for pi-coding-agent. Plans, state, MANDATORY rules, self-review, multi-question picker. One npm install, zero config. LLM drives the workflow inline via the soly_workflow tool — no external subagent plugin.

Multi-provider notifications when Pi settles, with optional inactivity reminders.

A subagent tool supporting context compaction and nested subagents — user-customizable models (with round-robin) and tool policies, extensible via hooks.

Questionnaire-guided design interview command and persistence tools for Pi.

Safe git-worktree management for pi: create/list/merge/remove with safety rails, no shell interpolation, Windows-first, zero tmux

Fuzzy file finding and indexed content search for pi: a Rust trigram index that survives compaction, with a zero-dependency TypeScript fallback

Pi extension for fan-out code review: N parallel reviewer subagents + an inline gate via subagent workflowScript

pi-CAW: reusable Roles, Workflows and journaled Runs for Pi hosts

pi coding agent extension: switch between projects under a configurable base directory via /project

Interactive active-tool manager for Pi.

Pi model provider that auto-discovers models from a Lemonade Server and integrates them into Pi via the server's OpenAI-compatible API

Pi TUI configuration and an on-demand Skill/CLI for image generation.

Free artifact publishing for Pi: share any HTML or Markdown file your agent writes as a link anyone can open. One tool call, no deploy, no hosting.

Delegate coding-agent work by spawning other agents as herdr tabs.

Precompute Pi compaction summaries in the background and reuse them when compaction starts.

Compact oversized Pi tool results with transparent previews and spill files.

Independently maintained MCP client for Pi with a programmatic source lifecycle

Interactive ask_user_question tool for pi-coding-agent with searchable split-pane selection UI, multi-select, and freeform input

Pi extension that registers avenor agent runs as tools in your Pi session. When you need a sub-agent to handle a well-defined task — write code, run tests, search a codebase — avenor operates in the background while you keep your session focused.