pi-producthunt

内容来源：README.md（说明文档） · 原始地址 · 查看安装指南

原始内容

# Pi Product Hunt

[![CI](https://github.com/eiei114/pi-producthunt/actions/workflows/ci.yml/badge.svg)](https://github.com/eiei114/pi-producthunt/actions/workflows/ci.yml)
[![Publish](https://github.com/eiei114/pi-producthunt/actions/workflows/publish.yml/badge.svg)](https://github.com/eiei114/pi-producthunt/actions/workflows/publish.yml)
[![npm version](https://img.shields.io/npm/v/pi-producthunt.svg)](https://www.npmjs.com/package/pi-producthunt)
[![npm downloads](https://img.shields.io/npm/dm/pi-producthunt.svg)](https://www.npmjs.com/package/pi-producthunt)
[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
[![Pi package](https://img.shields.io/badge/pi-package-purple.svg)](https://pi.dev/packages)
[![Trusted Publishing](https://img.shields.io/badge/npm-Trusted%20Publishing-blue.svg)](docs/release.md)
<a href="https://buymeacoffee.com/ekawano114m"><img src="https://cdn.buymeacoffee.com/buttons/v2/default-yellow.png" alt="Buy Me A Coffee" width="217" height="60"></a>

> Product Hunt research and digest workflows inside Pi.

## What this is

Pi Product Hunt is a read-only Pi extension package for Product Hunt market research. It calls the Product Hunt GraphQL API directly, adds interactive `/producthunt:*` commands for humans, and exposes structured tools for agents to gather launches, post details, comments, and digest-ready research notes.

## Features

- Daily Product Hunt launch scans.
- Product/post search for competitor and trend research.
- Post detail and comment collection for user-reaction analysis.
- Digest-ready Markdown with sections for signals, reactions, and topic watchlists.
- Compact topic watchlists with bounded rationale for products worth revisiting.
- Bounded markdown product cards for pasting launches into research notes.
- Persistent login that stores your Product Hunt token in the Pi agent directory.
- Agent tools with typed parameters for autonomous Product Hunt research.

## Install

Install the published npm package with Pi:

```bash
pi install npm:pi-producthunt
```

Pin a specific version when you want reproducible installs:

```bash
pi install npm:pi-producthunt@0.3.2
```

Install into the current project instead of your user Pi settings:

```bash
pi install npm:pi-producthunt -l
```

Or install from GitHub:

```bash
pi install git:github.com/eiei114/pi-producthunt
```

Try without installing permanently:

```bash
pi -e npm:pi-producthunt
```

For local development from a checkout:

```bash
pi -e .
```

## Quick start

1. Install the package:

```bash
pi install npm:pi-producthunt
```

2. Authenticate with Product Hunt:

```txt
/producthunt:login
```

3. Fetch today's launches:

```txt
/producthunt:today
```

Or search for a topic:

```txt
/producthunt:search
```

Agents can fetch Product Hunt data directly with typed tools:

```txt
producthunt_get_posts({ limit: 5 })
producthunt_search_posts({ query: "AI coding agent", limit: 10 })
```

## Authentication

Product Hunt access comes from one of two sources:

1. **Environment variable** `PRODUCTHUNT_ACCESS_TOKEN`
2. **Stored login** written by `/producthunt:login` to `~/.pi/agent/pi-producthunt-auth.json`

**Precedence:** when both are set, the environment variable wins. The stored login file is ignored until you unset the environment variable or start Pi without it.

Use the interactive login command:

```txt
/producthunt:login
```

Remove the stored token with:

```txt
/producthunt:logout
```

`/producthunt:logout` only clears the stored login file. It does not change `PRODUCTHUNT_ACCESS_TOKEN` in your shell.

Set a token through the environment when you prefer shell-based configuration:

```bash
export PRODUCTHUNT_ACCESS_TOKEN=...
```

Check auth without exposing secrets:

```txt
/producthunt:status
```

`/producthunt:status` reports the active token source (environment vs stored login), validates the token against the Product Hunt API when possible, and prints concise recovery steps for common failures:

- missing token
- unreadable or invalid stored login file
- rejected or invalid API token
- transient validation errors

Token values are always redacted from status output, logs, and error messages.

## Commands

Commands are human-facing and require no fixed inline arguments. If input is needed, Pi asks interactively.

```txt
/producthunt:status
/producthunt:login
/producthunt:logout
/producthunt:today
/producthunt:search
/producthunt:post
/producthunt:comments
/producthunt:digest
/producthunt:research
/producthunt:watchlist
/producthunt:cards
```

Example flows:

```txt
/producthunt:today      # today's launch list
/producthunt:search     # asks for a search query
/producthunt:post       # asks for slug, ID, or URL
/producthunt:comments   # asks for slug, ID, or URL
/producthunt:digest     # asks for today / yesterday / custom date
/producthunt:research   # asks for a research topic
/producthunt:watchlist  # asks for a topic and returns a compact revisit list
/producthunt:cards      # asks for a topic and returns paste-ready product cards
```

## Agent tools

Agents can call these typed tools directly:

```txt
producthunt_status
producthunt_get_posts
producthunt_search_posts
producthunt_get_post
producthunt_get_post_comments
producthunt_research_topic
producthunt_topic_watchlist
producthunt_research_product_cards
producthunt_digest
```

Examples:

```txt
producthunt_search_posts({ query: "AI coding agent", limit: 10 })
producthunt_get_post({ ref: "example-product-slug" })
producthunt_get_post_comments({ ref: "example-product-slug", limit: 10 })
producthunt_digest({ date: "2026-06-01", limit: 10 })
producthunt_topic_watchlist({ query: "AI coding agent", limit: 5 })
producthunt_research_product_cards({ query: "AI coding agent", limit: 5 })
```

See [`docs/watchlist.md`](https://github.com/eiei114/pi-producthunt/blob/HEAD/docs/watchlist.md) for when to use a watchlist vs a full digest or research pack.
See [`docs/research-pack.md`](https://github.com/eiei114/pi-producthunt/blob/HEAD/docs/research-pack.md) for when to prefer product cards over raw tool output.

## Package contents

| Path | Purpose |
|---|---|
| `extensions/` | Pi extension entrypoint and command/tool registration |
| `lib/` | Product Hunt API client, auth store, formatters, schemas, helpers |
| `docs/` | Release notes, usage examples, and watchlist guidance |

## Development

```bash
npm install
npm run ci
```

`npm run ci` runs:

- TypeScript typecheck
- Node tests
- `npm pack --dry-run`

## Release

This package is set up for npm Trusted Publishing, so no `NPM_TOKEN` is required.

```bash
npm version patch
git push
```

See [`docs/release.md`](https://github.com/eiei114/pi-producthunt/blob/HEAD/docs/release.md) for setup details.

## Security

Pi packages execute with your local permissions. Review source before installing third-party packages.

Product Hunt tokens are never committed by this package. `/producthunt:login` stores the token locally in `~/.pi/agent/pi-producthunt-auth.json`; `/producthunt:logout` deletes that stored file. `PRODUCTHUNT_ACCESS_TOKEN` is never modified by logout.

For vulnerability reporting, see [`SECURITY.md`](https://github.com/eiei114/pi-producthunt/blob/HEAD/SECURITY.md).

## Links

- npm: https://www.npmjs.com/package/pi-producthunt
- GitHub: https://github.com/eiei114/pi-producthunt
- Issues: https://github.com/eiei114/pi-producthunt/issues

## License

MIT

Pi Product Hunt

Product Hunt research and digest workflows inside Pi.

What this is

Pi Product Hunt is a read-only Pi extension package for Product Hunt market research. It calls the Product Hunt GraphQL API directly, adds interactive /producthunt:* commands for humans, and exposes structured tools for agents to gather launches, post details, comments, and digest-ready research notes.

Features

Daily Product Hunt launch scans.
Product/post search for competitor and trend research.
Post detail and comment collection for user-reaction analysis.
Digest-ready Markdown with sections for signals, reactions, and topic watchlists.
Compact topic watchlists with bounded rationale for products worth revisiting.
Bounded markdown product cards for pasting launches into research notes.
Persistent login that stores your Product Hunt token in the Pi agent directory.
Agent tools with typed parameters for autonomous Product Hunt research.

Install

Install the published npm package with Pi:

pi install npm:pi-producthunt

Pin a specific version when you want reproducible installs:

pi install npm:pi-producthunt@0.3.2

Install into the current project instead of your user Pi settings:

pi install npm:pi-producthunt -l

Or install from GitHub:

pi install git:github.com/eiei114/pi-producthunt

Try without installing permanently:

pi -e npm:pi-producthunt

For local development from a checkout:

pi -e .

Quick start

Install the package:

pi install npm:pi-producthunt

Authenticate with Product Hunt:

/producthunt:login

Fetch today's launches:

/producthunt:today

Or search for a topic:

/producthunt:search

Agents can fetch Product Hunt data directly with typed tools:

producthunt_get_posts({ limit: 5 })
producthunt_search_posts({ query: "AI coding agent", limit: 10 })

Authentication

Product Hunt access comes from one of two sources:

Environment variable PRODUCTHUNT_ACCESS_TOKEN
Stored login written by /producthunt:login to ~/.pi/agent/pi-producthunt-auth.json

Precedence: when both are set, the environment variable wins. The stored login file is ignored until you unset the environment variable or start Pi without it.

Use the interactive login command:

/producthunt:login

Remove the stored token with:

/producthunt:logout

/producthunt:logout only clears the stored login file. It does not change PRODUCTHUNT_ACCESS_TOKEN in your shell.

Set a token through the environment when you prefer shell-based configuration:

export PRODUCTHUNT_ACCESS_TOKEN=...

Check auth without exposing secrets:

/producthunt:status

/producthunt:status reports the active token source (environment vs stored login), validates the token against the Product Hunt API when possible, and prints concise recovery steps for common failures:

missing token
unreadable or invalid stored login file
rejected or invalid API token
transient validation errors

Token values are always redacted from status output, logs, and error messages.

Commands

Commands are human-facing and require no fixed inline arguments. If input is needed, Pi asks interactively.

/producthunt:status
/producthunt:login
/producthunt:logout
/producthunt:today
/producthunt:search
/producthunt:post
/producthunt:comments
/producthunt:digest
/producthunt:research
/producthunt:watchlist
/producthunt:cards

Example flows:

/producthunt:today      # today's launch list
/producthunt:search     # asks for a search query
/producthunt:post       # asks for slug, ID, or URL
/producthunt:comments   # asks for slug, ID, or URL
/producthunt:digest     # asks for today / yesterday / custom date
/producthunt:research   # asks for a research topic
/producthunt:watchlist  # asks for a topic and returns a compact revisit list
/producthunt:cards      # asks for a topic and returns paste-ready product cards

Agent tools

Agents can call these typed tools directly:

producthunt_status
producthunt_get_posts
producthunt_search_posts
producthunt_get_post
producthunt_get_post_comments
producthunt_research_topic
producthunt_topic_watchlist
producthunt_research_product_cards
producthunt_digest

Examples:

producthunt_search_posts({ query: "AI coding agent", limit: 10 })
producthunt_get_post({ ref: "example-product-slug" })
producthunt_get_post_comments({ ref: "example-product-slug", limit: 10 })
producthunt_digest({ date: "2026-06-01", limit: 10 })
producthunt_topic_watchlist({ query: "AI coding agent", limit: 5 })
producthunt_research_product_cards({ query: "AI coding agent", limit: 5 })

See docs/watchlist.md for when to use a watchlist vs a full digest or research pack. See docs/research-pack.md for when to prefer product cards over raw tool output.

Package contents

Path	Purpose
`extensions/`	Pi extension entrypoint and command/tool registration
`lib/`	Product Hunt API client, auth store, formatters, schemas, helpers
`docs/`	Release notes, usage examples, and watchlist guidance

Development

npm install
npm run ci

npm run ci runs:

TypeScript typecheck
Node tests
npm pack --dry-run

Release

This package is set up for npm Trusted Publishing, so no NPM_TOKEN is required.

npm version patch
git push

See docs/release.md for setup details.

Security

Pi packages execute with your local permissions. Review source before installing third-party packages.

Product Hunt tokens are never committed by this package. /producthunt:login stores the token locally in ~/.pi/agent/pi-producthunt-auth.json; /producthunt:logout deletes that stored file. PRODUCTHUNT_ACCESS_TOKEN is never modified by logout.

For vulnerability reporting, see SECURITY.md.

License

MIT