deemwar
Working capabilities you can run, pulled from real problems. Each one leads with the job it does for you, ships as self-contained code you run locally in a sitting, and drops into your own stack. Free, MIT. Engineering-first — no slideware.
Drop one wrapper around your LLM client and get a hard cap that actually holds when workers run concurrently, a per-agent spend breakdown (who burned the most, in dollars), and an early pause when a loop goes runaway. Runnable — node report.js prints the spend report. No network, zero deps.
More examples land here as we hit the problems worth writing up. Each is judged on one thing: does the code run and hold up.