npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@teelicht/pi-superagents

v0.9.1

Published

Pi extension for Superpowers workflows: recon, research, implementation, review, and debug

Downloads

680

Readme

@teelicht/pi-superagents

Pi agent-harness extension to support superpowers workflows using subagents.

Features

  • Superpowers Workflow: Proven pipeline for robust AI-assisted development.
  • Role-Specific Agents: Thin agents-layer for every phase of the development lifecycle.
  • Model Tiers & Command Settings: Abstract model selection (cheap, balanced, max) for each agent. One model and optional thinking level can be configured per tier. Custom tiers are possible. Models, tier thinking, and command-scoped behavior toggles can be changed through the /sp-settings TUI; its model picker supports type-to-search filtering, including q, and scrolls through large authenticated model lists. Slash-command metadata lives in entrypoint agent frontmatter.
  • Plannotator Integration: Optional event bridge to Plannotator for browser-based spec/plan review and approval (setting).
  • Worktree Isolation: Optional git worktree creation for parallel tasks to prevent filesystem conflicts. Worktree-backed parallel children are joined before cleanup; worktree policy is unchanged.
  • Entrypoint Lifecycle Skills: Superpowers entrypoint agents inject root lifecycle skills (verification, review-feedback, branch finishing) defined in frontmatter. .
  • Subagent Extension & Tool Defaults: Subagents run with implicit Pi extension discovery disabled by default; configure superagents.extensions with local paths or Pi -e source specs, and superagents.tools with shared tool names or tool extension paths. The bundled defaults provide the common read-only tools globally so agent frontmatter only lists role-specific extras.

Installation

pi install npm:@teelicht/pi-superagents

[!NOTE] This tool requires the superpowers skills to be installed. Easy installation via https://skills.sh/obra/superpowers.

On install, pi-superagents creates config.json from the bundled defaults:

~/.pi/agent/extensions/subagent/config.json

To remove:

pi remove npm:@teelicht/pi-superagents

Execution Model

Subagent execution is synchronous and blocking. The subagent tool does not accept async, wait, collect, or cancel parameters. Lifecycle tools (subagent_done, caller_ping) are registered as internal child-only tools through policy; they are not general-purpose delegation tools.

Integrated Slash Commands

Superpowers slash commands are registered from interactive entrypoint agent frontmatter, not generated from config.json. config.json only changes runtime behavior for commands that already have a matching entrypoint agent. Use superagents.tools to append shared tool names or tool extension paths to every subagent without repeating them in each agent frontmatter file.

| Command | Description | | ----------------------- | --------------------------------------------------------------------------- | | /sp-brainstorm <task> | Brainstorm a task and save a spec, optionally review it with Plannotator UI | | /sp-plan <task> | Plan a task with optional Plannotator plan review | | /sp-implement <task> | Run an implementation task through the Superpowers flow | | /subagents-status | Open active and recent subagent run status, including runtime-confirmed models, thinking levels, and resolved skills | | /sp-settings | Open superagents settings |

Custom Commands

Create custom slash commands by adding an interactive entrypoint agent markdown file (e.g., ~/.pi/agent/agents/sp-mycommand.md) with the appropriate frontmatter. Optional behavior flags (e.g., useSubagents, usePlannotator) can be set in config.json under superagents.commands.<name>. Config blocks alone do not register commands.

See Configuration for the agent frontmatter schema and behavior flag reference.

Agents

The /sp-implement command activates a structured workflow for task execution with an interactive entrypoint agent, role-specific headless agents, model tiers, and built-in quality gates. The bundled agents/sp-implement.md entrypoint injects root lifecycle skills for verification, review-feedback handling, and branch finishing. The bundled sp-debug role injects systematic-debugging when delegated.

Subagent execution remains conservative and synchronous for ordinary Superpowers workflows. There is intentionally no user-facing async or blocking switch in agent frontmatter, config, or tool parameters. Internal result ownership prevents duplicate delivery and lifecycle sidecars let child agents report intentional completion or a parent-help request without changing the normal delegation flow.

  1. Recon (sp-recon): Initial codebase analysis and context gathering.
  2. Research (sp-research): Deep dive into specific APIs, libraries, or logic.
  3. Implementation (sp-implementer): Code changes guided by test-driven development (optional).
  4. Review (sp-code-review): Automated review of changes against project standards.
  5. Debug (sp-debug): Root cause analysis and fix verification for regressions.

Subagent-driven development keeps implementer and reviewer reports inline in the Pi conversation. Bounded roles default to lineage-only - they see a curated work brief rather than the full parent conversation history. The runtime does not create repo-root packet files such as implementer-report.md, spec-review.md, code-review.md, debug-brief.md, or task-brief.md; those names are ignored if an older prompt or manual run creates them.

Run history is persisted at ~/.pi/agent/run-history.jsonl for /subagents-status. Inline subagent rows and the status overlay show the model reported by the child Pi execution loop and, when available, the effective thinking level used for that run. Set PI_SUPERAGENTS_RUN_HISTORY_PATH to isolate that file for tests or sandboxed sessions.

Configuration & Documentation

Credits