Skip to content

Instantly share code, notes, and snippets.

@swapp1990
swapp1990 / the-bottle-lantern.md
Created February 15, 2026 07:18
Story Agent Test: The Bottle Lantern (mystery, 3 chapters)

The Bottle Lantern

On a remote island where the sea keeps its own counsel, lighthouse keeper Thomas finds a message in a bottle that shouldn’t have survived the decades. The note points to a long-buried shipwreck, a missing treasure, and a woman named Eleanor whose past is knotted to the island’s darkest night. As Thomas follows the trail through salt-stained records and whispered local lore, the truth surfaces with a price neither of them expected to pay.


Chapter 1: A Note That Wouldn’t Sink

Rain had a way of making the island feel like it was being erased in real time. It came down in thin, relentless lines that blurred the horizon into a smear of pewter and turned every surface honest—stone showed its slick black heart, grass flattened into a dark pelt, and the lighthouse glass wore a film of salt like a guilty sheen.

@swapp1990
swapp1990 / story-agent-plan.md
Created February 15, 2026 06:16
FanFlix Story Writing Agent - MVP Plan

Story Writing Agent — MVP Plan

1. Problem Statement

FanFlix needs a way to generate complete short stories autonomously from a premise or story idea. The existing LMWFY writing system only handles paragraph-level operations (continue, rewrite, insert) with a human in the loop. There is no capability to produce a full, coherent multi-chapter story end-to-end without manual intervention at every step.

2. Solution Overview

Add a new story_agent product to the LMWFY backend that orchestrates complete story generation. A single API call with a premise, genre, and optional structure hints triggers an agent that plans the story (outline, chapters, character arcs), then writes each chapter sequentially — passing accumulated context forward so the narrative stays coherent. The client polls one job ID and receives chapters as they complete, with the full story available at the end.

@swapp1990
swapp1990 / playwright-mcp-setup-plan.md
Created February 15, 2026 01:29
Playwright MCP Setup Plan for Telegram→Claude Browser Automation

Playwright MCP Setup for Telegram→Claude Browser Automation

1. Problem Statement

Claude in Chrome has persistent auth/connection issues and isn't designed for automated tasks triggered via Telegram /run commands. We need reliable browser automation that Claude Code can use natively — both in interactive sessions (where Swap watches) and in autonomous /run tasks (headless, in tmux). Currently there's no way to do browser actions like checking App Store keyword rankings through the Telegram→MacBook→Claude pipeline.

2. Solution Overview

Install Microsoft's Playwright MCP as a global Claude Code tool. Configure it for two modes: headless (for /run tasks in tmux) and headed (for interactive sessions). Set up a persistent browser profile so logged-in sessions survive across runs. Validate with an end-to-end test that goes from a Telegram /run command all the way through to browser automation and result delivery back to Telegram.

@swapp1990
swapp1990 / tiktok-trend-finder-plan.md
Created February 12, 2026 22:55
TikTok Trend Finder - Plan

TikTok Trend Finder — Plan

1. Problem Statement

There's no easy, free way to find the top trending TikTok videos for a specific keyword (like "iOS app demo") on a given day, with direct links. TikTok's own tools only filter by broad categories, not custom search terms. Swap needs this to research trending content in specific niches for marketing and content strategy.

2. Solution Overview

A Python CLI tool called tiktok-trend-finder that takes a search query, calls the TokInsight API (free tier, 2000 calls/day, no credit card), and returns the top N trending TikTok videos sorted by view count — each with title, views, likes, and a direct TikTok link. If TokInsight search is insufficient, the tool swaps to davidteather/TikTok-Api (Playwright-based) as a fallback.

@swapp1990
swapp1990 / test-context-prompt.md
Created February 12, 2026 22:26
Thread context injection test — Marketing (1319) prompt

SESSION_TYPE: run RESPONSE_CHANNEL: telegram-thread-1319 INITIATED_BY: swap

You are executing an autonomous task.

Marketing (Thread 1319)

Purpose: Marketing campaigns, social media content, playbooks, client content strategy

@swapp1990
swapp1990 / discuss-1770906513196-turn-2.md
Created February 12, 2026 14:38
Discussion reply from discuss-1770906513196

Perfect! This clarifies everything. You're building an AI-agent-operated screenshot generator - way more interesting than just a script. Here's the architecture:

System Architecture

Two-Phase Pipeline

Phase 1: CAPTURE (new - iOS Simulator automation)

  • Launch simulator for the project
  • Navigate through key app screens
  • Capture raw screenshots
@swapp1990
swapp1990 / discuss-1770906250958-turn-2.md
Created February 12, 2026 14:36
Discussion reply from discuss-1770906250958
@swapp1990
swapp1990 / upwork-categories-2026-report.md
Created February 12, 2026 14:35
Upwork in 2026: Top Job Categories & Morgan Overholt's K Success Story - Deep Research Report

Upwork in 2026: Top Job Categories & Successful Freelancer Profile

Research Date: February 12, 2026 Focus: Deep dive into Upwork's job market structure and real-world success story


Executive Summary

Upwork's 2026 marketplace is dominated by Web, Mobile & Software Development (34%), followed by Writing (18%), Admin & Support (11%), and Design & Creative (9%). However, the explosive trend is AI-augmented skills growing 109% YoY, with specific AI categories like video generation (+329%) and AI integration (+178%) leading the charge. This report includes a detailed profile of Morgan Overholt, a graphic designer who has earned $700,000+ on Upwork and exemplifies the modern successful freelancer.

@swapp1990
swapp1990 / software-gig-work-2026-report.md
Created February 12, 2026 14:26
The State of Software Gig Work in 2026: Deep Research Report on AI's Impact on Freelance Development

The State of Software Gig Work in 2026: A Deep Research Report

Research Date: February 12, 2026 Topic: The transformation of software development gig work in the AI era


Executive Summary

The software development gig economy in 2026 has undergone a profound transformation driven by AI coding tools. The market has split into two distinct tiers: elite developers who leverage AI to multiply their productivity are thriving, while entry-level and routine developers face significant displacement. Simple website and app development work has become largely commoditized, with AI tools enabling non-developers to build functional sites. However, architecture, system design, AI/ML integration, security, and specialized development remain in high demand.

@swapp1990
swapp1990 / discuss-1770768048101-turn-4.md
Created February 11, 2026 00:44
Discussion reply from discuss-1770768048101

The submission form is behind a login, but based on RevenueCat's official guidelines + standard Devpost structure, here's what they're asking for:

Standard Devpost Fields (All Hackathons)

Basic Info:

  • Project Name - "Quest Planner"
  • Tagline (short, catchy) - "Turn your dreams into daily micro-actions"
  • Which Creator Brief? - Dropdown: Select "Gabby Beckford (@packslight)"

Assets: