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

psypi

v0.4.1

Published

A standalone, evidence-grounded social-science research agent — a conversation-driven agent that bundles the Pi runtime with psypi's research contracts, evidence gates, and workflows

Downloads

121

Readme

psypi

面向社会科学研究者的 Pi 科研智能体层。

psypi 的目标不是再造一个重量级 Agent,而是在官方 Pi 之上提供少量、可审计、可组合的科研契约:研究项目状态、Claim-Evidence-Provenance 账本、完整性门禁、可恢复工作流、受控记忆和按需技能。统计计算交给成熟库或外接 MCP;Pi 的会话、模型、Skill、Package 和 TUI/RPC 能力继续由 Pi 提供。

当前仓库已完成第一条 M1/MVP-0 可运行切片,并落地 M2 的技能注册与受限多智能体编排基础、M3 的只读并行契约、真实 stdio MCP transport(argv/显式 env/超时隔离)、Skill 恶意正文 preflight、许可证/SBOM 门与 M4 的 literature-review/analysis-delegation/writing-review 三个 workflow 包。官方 Pi 的 AgentSessionSessionManagerDefaultResourceLoader、模型网关和严格 JSONL RPC 已由 adapter 接入;完成项目级 package 安装后提供原生只读 /agents 多智能体入口,面板扩展可选安装。联网检索、统计计算和并行写入仍按规划后置。先阅读:

代号:psypi。当前可用命令:

pnpm install
pnpm build
pnpm exec pi install -l .       # 一次性把 psypi 作为项目级 Pi package 启用
pnpm exec pi --approve           # 运行 Pi 时信任已审查的项目 package
node dist/src/cli.js init "研究问题" --paradigm survey-observational
node dist/src/cli.js evidence add notes/source.md --level user
node dist/src/cli.js brief
node dist/src/cli.js workflow literature-review
node dist/src/cli.js serve --port 8787
  node dist/src/cli.js agents
  node dist/src/cli.js install claude-code --yes  # 当前 agent 全局命令仍为 discover-only
  node dist/src/cli.js import claude-code --yes
  node dist/src/cli.js shell
  /skills                         # 查看推荐 Skill 的启用状态
  /skills enable <skill-id>       # 启用项目级 Skill
  /skills disable <skill-id>      # 禁用项目级 Skill
  /mcp                            # 查看推荐 MCP 的启用状态
  /mcp enable <mcp-id>            # 启用项目级 MCP
  /mcp disable <mcp-id>           # 禁用项目级 MCP
  /install skill|mcp <id>         # 生成安装预检计划
  pnpm exec pi --extension ./dist/src/extension.js --approve  # 未安装 package 时的临时入口
  pi --extension ./dist/src/panel/extension.js                # 可选:在 Pi 中启停 /panel

pi-workbench 中使用时,先在研究项目根目录安装 psypi package:

pi install -l F:\Projects\psypi\psypi
pi-workbench

workbench 启动的每个 Pi 会话都会加载 psypi 的 /research/brief/agents/panel 命令。在 Pi 中执行 /panel 后,用浏览器打开提示的 http://127.0.0.1:8787 地址即可查看当前项目的只读运行状态。

briefworkflow 在证据门禁失败时退出码为 2,并保留 manifest 与 HANDOFF,不生成貌似完成的正文。serve 启动本机面板:运行事实、Agent/Skill 和模型目录为只读;模型配置表单是唯一明确的本地写入入口,会将 provider 元数据写入 Pi models.json,API key 交给 Pi auth.json 管理。

内置 agent 下载器/安装器:agents/scan 只读识别本机其他智能体(Claude Code、OpenAI Codex、Gemini CLI、opencode、Aider、Cursor、Windsurf、Continue、Orca、Copilot CLI)的配置与技能目录,绝不读取凭据内容;安装计划必须有来源、版本、artifact SHA、许可证/依赖审计和 SBOM,缺任何一项都会 blocked。当前 agent catalog 只有全局 npm/pipx 命令,尚未绑定可验证 staging artifact,因此即使人工确认也保持 agent-staging-required,不会执行全局安装;import <id> --yes 将其他 agent 的允许 Skill 文件以来源/SHA-256 溯源导入 .psypi/imports/shell 启动 ink 交互式表格 TUI(↑/↓ 选择、i 安装、m 导入、r 重扫、q 退出)。

Pi 集成分两层:完成一次 pi install -l . 后,项目级 package 默认加载 src/extension.ts,注册研究命令和原生只读 /agents;未安装时可用上面的 --extension 临时入口。worker 通过独立 Pi RPC 子进程、read/grep/find/ls 工具和结构化 WorkerReport 回传,但进程边界仍不是 OS sandbox。src/panel/extension.ts 是可选扩展,使用 /panel [--port 8787] 启停仅绑定 127.0.0.1 的只读可视化界面。面板的 Agents/Skills 下载页只展示发现结果、来源、版本、许可证、SHA、风险和审批前计划;模型页读取 Pi 本地模型注册表,并在没有对应注册项时显示 psypi 的 DeepSeek 模板元数据。页面只显示环境变量名称,不接收或显示 API key;configured 只表示环境变量存在,不代表实际认证成功。模型适配沿用 Pi models.json/registerProvider 契约,DeepSeek helper 只输出 $DEEPSEEK_API_KEY 引用。