Plugin registry

Plugins for pansy, OpenCode and Pi, ranked by what npm, GitHub and their users say.

  1. 2401
    P

    Declarations of a source file with line spans, so a read fetches one body instead of the whole file.

  2. 2402
    P

    Replaces the default footer with a Claude Code-style layout: model, context usage, cost, branch, worktree.

  3. 2403
    P

    Pi sub-agent orchestration package for Yishan.

  4. 2404
    P

    Worktrunk commands, session placement, status, and agent tools for pi

  5. 2405
    P

    Steward — the llama.cpp control panel for Pi. A local browser dashboard for service control, resident models, host health, and streamed logs.

  6. 2406
    P

    Pi extension: `#` + model name autocomplete

  7. 2407
    P

    Autonomous experiment loop that costs nothing until you type /autoresearch: measure, decide, commit or revert, with scope and benchmark guards.

  8. 2408
    P

    Native Pi extension for Future AGI turn, model and tool tracing

  9. 2409
    P

    Injects dynamic system-reminder messages before each LLM call from a shared generator registry.

  10. 2410
    P

    Pi extension: DeepSeek Search & Fetch tools for real-time web search and page extraction, reusing your existing DeepSeek API key

  11. 2411
    P

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

  12. 2412
    P

    maestria extension for the Pi coding agent

  13. 2413
    P

    Background subagents on a pi or Claude Code backend, with fire-and-forget spawn and deferred result delivery.

  14. 2414
    P

    Claude Code style TaskCreate, TaskGet, TaskList, and TaskUpdate tools for pi, a more faithful port of Claude Code's harness than @tintinweb/pi-tasks.

  15. 2415
    P

    Pi TUI overlay wizard: manage models.json providers (add a relay or gateway, fetch /models, edit capabilities) and review token usage by provider/model

  16. 2416
    P

    Reads what the sessions repeated, writes the skill or rule that would have prevented it, and takes it back out when the rate does not drop.

  17. 2417
    P

    One bg tool for background processes, woken on exit, with cursor reads that never re-read bytes.

  18. 2418
    P

    Pi extension for social research

  19. 2419
    P

    Wool provider extension for Trae (ByteDance) — custom SSE + Cloud-IDE-JWT + fixed header family, storage.json container credential extraction (SOLO free-chat channel).

  20. 2420
    P

    Enhanced Qwen Token Plan CN provider for pi. Speaks the OpenAI Responses API (not chat-completions) to activate the platform's server-side Harness tools — web search, code interpreter, web extraction, and image search — that the built-in qwen-token-plan-c

  21. 2421
    P

    Pi extension: detects and interrupts infinite thinking-block and tool-call loops in real time before they exhaust your context window.

  22. 2422
    P

    Token-budgeted repo mapping for agent sessions: foveated heat diffusion over a cross-language code graph, with progressive disclosure.

  23. 2423
    P

    Keep retrying failed provider requests that Pi's built-in retry gives up on

  24. 2424
    S

    Local image-code knowledge library with native macOS and local Web clients, plus MCP and Skill for Pi, DeepSeek Harness (dsh), Claude, Codex, Cursor, and Wisp.

  25. 2425
    P

    Pi harness package for companion readiness. Product workflow lives in Grupo Ilao engineering skills.

  26. 2426
    P

    Pi extension: quick model + thinking switch buttons

  27. 2427
    P

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

  28. 2428
    P

    Detailed guidance for creating Agent Skills (Pi-compatible Agent Skills format).

  29. 2429
    F

    Display-only rich Markdown, equations, diagrams and private previews for Pi with runtime compatibility checks.

  30. 2430
    P

    Pi's native HTML export with live theme colors, offline LaTeX, equation copying, sharing previews, and print/PDF

  31. 2431
    P

    Give your Pi agent a counsel tool: a stronger model it consults before committing to an approach, when stuck, or before declaring done.

  32. 2432
    P

    Doom-style Pi status bar with a high-resolution image face, universal quadrant-block fallback, live context and cost stats, and test-health controls

  33. 2433
    P

    Multi-server Forgejo workflows, dashboard, reviews, and Actions tools for Pi

  34. 2434
    P

    General-purpose permission system for pi tools, handling permissions for bash and file tools with extensible matchers for custom tools.

  35. 2435
    P

    Pi integration for Lab Agent Workbench

  36. 2436
    P

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

  37. 2437
    D

    Snapshot-bound hashline read tool for Pi and DoomPi.

  38. 2438
    P

    Pi extension: web_search + web_fetch with zero-config Exa MCP free search, keyless Bing, self-hosted SearXNG, and pluggable providers (Bocha, Tavily, Exa, Brave, Jina, Firecrawl).

  39. 2439
    P

    Convert Slack and GitHub emoji shortcodes in Pi.

  40. 2440
    C

    Pi extension that gathers git diff and project context to draft or create clear Conventional Commit messages.

  41. 2441
    P

    Plan on the frontier model, execute on the cheap one. The session model swaps the moment the first edit lands.

  42. 2442
    P

    pi extension: switch a session between pi's stock system prompt, a trimmed variant, and a bundled or custom persona prompt - for prompt-engineering experiments

  43. 2443
    P

    A standalone pi extension that adds /annotate-git-diff, a native Glimpse UI for annotating git diffs and sending explicit feedback directly to the agent; close recovery keeps unsent drafts in the editor.

  44. 2444
    E

    Exit command extension for pi-agent.

  45. 2445
    P

    编辑器上方的一行状态条(模型、思考、目录、分支、上下文、缓存、费用)和下方的上一条提示,不替换编辑器

  46. 2446
    P

    Streaming-first rich Markdown, code and math rendering for the pi TUI, with formula images in Kitty graphics terminals (display only; never touches the LLM context).

  47. 2447
    P

    dsh-style thinking display for pi: one-line header with token count (configurable decimals, default 2); collapsed thinking shows a single-line content preview

  48. 2448
    P

    Render Mermaid diagrams from the conversation in the browser.

  49. 2449
    A

    Amber Protocol: the governed execution boundary between AI agents and real systems — repository-local plans, gates, approvals, evidence, and handoff.

  50. 2450
    P

    Subagents for pi that never lose work and never do it twice. Crash-safe workflows, automatic recovery, and a live view just like the main agent.