Skip to content

CLAUDE.md

Latest commit

 

History

History
77 lines (53 loc) · 3.98 KB

CLAUDE.md

File metadata and controls

77 lines (53 loc) · 3.98 KB

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Overview

@bnbot/cli — CLI for BNBot. Subcommand format: bnbot <platform> <action>.

Published as @bnbot/cli on npm. Binary names: bnbot, bnbot-cli (both point to dist/index.js).

Build & Dev

npm run build          # tsc → dist/
npm run dev            # tsx src/index.ts (no build needed)
npm version patch      # bump version
npm publish            # publish to npm (runs build via prepublishOnly)

No test suite or linter configured.

Release

npm version patch      # or minor/major
npm publish            # auto-runs build via prepublishOnly
git add -A && git commit -m "chore: bump to vX.Y.Z"
git push

Then update global install: npm i -g @bnbot/cli --ignore-scripts

Note: postinstall script downloads Claude skill from bnbot.ai/skill.md. Use --ignore-scripts if network is unavailable.

Architecture

Two execution paths based on whether a platform needs browser authentication:

Browser scrapers (X, TikTok, YouTube, Instagram, etc.):
  CLI → WebSocket (ws://localhost:18900) → Chrome Extension → chrome.scripting → platform DOM

Public scrapers (HackerNews, Wikipedia, StackOverflow, BBC, etc.):
  CLI → direct HTTP fetch (no extension needed)

Key Modules

  • src/index.ts — Entry point. Builds the commander program with all subcommands. Also handles legacy kebab-case routing (bnbot post-tweet --text "Hi") for backward compat.
  • src/wsServer.ts — BnbotWsServer class. WebSocket server on port 18900 that bridges CLI ↔ Chrome Extension. Handles request-response matching via requestId, auto-retry on busy, auth token relay.
  • src/cli.ts — runCliTool / runCliAction functions. CLI client that connects to the WS server, sends an action, waits for result. The TOOL_MAP maps kebab-case CLI names to snake_case WS action types.
  • src/publicScrapers.ts — PUBLIC_SCRAPERS map. All direct-fetch scrapers (no extension). Uses undici ProxyAgent to respect http_proxy/https_proxy env vars.
  • src/commands/actions.ts — Commander action handlers for X platform (post, reply, like, scrape, navigate). Maps commander args → runCliAction calls.
  • src/commands/scraperActions.ts — Commander action handlers for browser-based platform scrapers (TikTok, YouTube, Reddit, etc.). Each sends a SCRAPER_* action type via WebSocket.
  • src/tools/mediaUtils.ts — Media file resolution utility (local files → data URLs, URL downloads). Used by post/reply commands for --media option.
  • src/auth.ts — Login flow. Tries clawmoney API key from ~/.clawmoney/config.yaml, falls back to email OTP via api.bnbot.ai.
  • src/setup.ts — bnbot setup command. Installs CLI globally + downloads Claude Code skill from bnbot.ai/skill.md to ~/.claude/commands/bnbot.md.

Adding a New Platform

Public scraper (no auth needed): Add a function to PUBLIC_SCRAPERS map in src/publicScrapers.ts, then add commander subcommands in src/index.ts.

Browser scraper (needs extension): Add action handler in src/commands/scraperActions.ts using the scrape() helper with a SCRAPER_* action type, then add commander subcommands in src/index.ts. The actual scraping logic lives in the Chrome Extension repo.

Related Projects

Project Path Description
CLI (this) /Users/jacklee/Projects/bnbot-cli @bnbot/cli npm package
Extension /Users/jacklee/Projects/bnbot-extension Chrome extension (browser scrapers)
Skill /Users/jacklee/Projects/bnbot /bnbot Claude Code skill
Frontend /Users/jacklee/Projects/bnbot-frontend bnbot.ai website
OpenCLI source /Users/jacklee/Projects/opencli-source opencli upstream reference

OpenCLI Sync

Every 3 weeks, diff against opencli upstream for API changes and new platforms. Sync log at /Users/jacklee/Projects/bnbot-extension/docs/OPENCLI_SYNC_LOG.md.