Text every World Cup final score to a friends' Signal group
A Claude Code-built agent checked ESPN every 30 minutes and sent each 2026 World Cup final score, written up in a persona, to a 3-person Signal group, with a code check on scores.
Done withClaude CodeClaude
- What they did
- Built with Claude Code in one evening. A macOS LaunchDaemon runs a Python script every 30 minutes. It pulls ESPN's public scoreboard, finds finished matches not yet sent, and has Claude Haiku write each result as the persona Claudia. A plain non-AI check confirms the score appears before sending through signal-cli. If the check fails, it sends the plain score line.
- How it went
- All 104 matches were sent over 39 days with no duplicates or send errors. 101 got persona write-ups and 3 fell back to plain lines. One shipped bug reported a penalty-shootout win as a 1–1 tie before it was fixed.
- Worth knowing
- Fetch yesterday's and today's ESPN scoreboards, and check match status so shootouts aren't read as draws. Cost was pennies, one short Haiku call per match.
Try it yourself with Claude Code
Build me a small agent that checks [ESPN or another scores site] every [30 minutes] during [tournament or season] and texts each final score for [team or league] to my [Signal/WhatsApp] group, written in the voice of [a persona, e.g. a dramatic sports announcer]. Add a code check so it only sends a score once the game is final and never sends the same result twice. Before you send anything to the group, show me a sample message and wait for my okay.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
Source: github.com · Oct 11, 2026 · Submitted by Bronson Elliott
Related use cases
Build a C compiler with a team of parallel agents
Anthropic's engineering write-up on tasking Opus 4.6 agent teams with building a C compiler.
Summarize 20 articles into a cited document
An agent extracted key insights from 20 articles, found common themes and produced a cited summary; the write-up estimates 18 minutes versus hours by hand.
Fill out 50 forms from public information
A user had Claude complete 50 individual work forms using public information, then proofread them and found them accurate.
Run an offline car assistant on a Raspberry Pi
A local Qwen model on a Pi reads car OBD data and the manual to answer car questions, controls AC and locks via cloud, and alerts other agents to a breakdown so they can look up train tickets.
Custom agent
Build Alexa routines to regulate a daily schedule
A TechRadar writer set Alexa routines to check mood and water intake at 11AM and 3PM and remind about lunch, saying it helped their day, with a warning against over-reliance.
Hand everyday errands to a local agent
A commenter lists tasks given to a trusted local agent: planning flights and hotels, converting music, finding DRM-free ebooks, designing 3D-printed holders, and tracking email appointments.
Custom agent
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.