📚 Browser Scripts Reference

Skill References

Browser Scripts Reference

Three browser console scripts for algorithm cultivation, with different trade-offs.

algorithmBuilder.js (Full-featured)

File: src/automation/algorithmBuilder.js
Requires: Paste src/automation/core.js first.

969-line browser automation with LLM-powered comment generation.

Features

  • 6 activity cycles: Search & Engage, Browse Home, Target Account, Explore, Search People, Own Profile
  • Weighted cycle selection with natural activity rotation
  • LLM-powered comment generation (OpenRouter or any OpenAI-compatible endpoint)
  • Session/break cycling for 24/7 operation
  • Peak/off-peak hour awareness
  • Per-session and per-day rate limits
  • Full state persistence via localStorage

Configuration

  • NICHE_CONFIG.KEYWORDS — search terms for your niche
  • NICHE_CONFIG.PERSONA — your niche identity (used for LLM persona)
  • NICHE_CONFIG.TARGET_ACCOUNTS — accounts to engage with
  • NICHE_CONFIG.BIO_KEYWORDS — only follow users matching these
  • LLM_CONFIG.ENABLED — enable AI-generated comments
  • LLM_CONFIG.API_KEY — your OpenRouter API key
  • LLM_CONFIG.MODEL — LLM model to use
  • BEHAVIOR — all timing, limits, and probability settings

Controls

  • window.stopAlgoBuilder() — stop after current action
  • window.algoStats() — print all-time stats
  • window.algoReset() — clear state and start fresh
  • window.algoConfig() — show current configuration

thoughtLeaderCultivator.js (Standalone)

File: scripts/thoughtLeaderCultivator.js

Zero dependencies. Paste directly into DevTools console on x.com.

Controls

  • stopCultivator() — stop after current action
  • cultivatorStatus() — print current stats
  • cultivatorReset() — clear all tracking data

algorithmTrainer.js (Multi-niche)

File: src/automation/algorithmTrainer.js
Requires: Paste src/automation/core.js first.

Multi-niche rotation, intensity presets, cycle-based architecture.

Controls

  • stopTrainer() — stop
  • trainerStatus() — status report
  • trainerReset() — clear data

Training Phases

All browser scripts cycle through these phases with randomized order:

  1. Search Top — search a niche term, scroll top results, engage
  2. Search Latest — same term, latest tab (catch fresh content)
  3. Search People — follow qualifying accounts in the niche
  4. Home Feed — scroll and engage to reinforce algorithm learning
  5. Influencer Visit — visit niche leaders' profiles with boosted engagement
  6. Own Profile — visit own profile (active user signal)
  7. Explore — browse trending topics (normalization behavior)
  8. Idle — dwell period simulating human rest

⚡ Explore XActions

100% free and open-source. No API keys, no fees, no signup.

Browse All Documentation