New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. The extreme engineering of aircraft windows(bbc.com ↗)
    discuss
  2. Computer Use for iPad(twitter.com/jamiepinheiro ↗)
    discuss
  3. Free Palestine(theguardian.com ↗)
    discuss
  4. Open HamClock Backend(ohb.works ↗)
    discuss
  5. The Roman telescope has enough gas for 22 years, double NASA's expectations(arstechnica.com ↗)
    discuss
  6. Breaking macOS Screen Time for fun and profit(dunkirk.sh ↗)
    discuss
  7. Apple Reference Image: A New Approach for Verified Photography(security.apple.com ↗)
    discuss
  8. Saudi Arabia discovers 110M tonnes of uranium-bearing ore in Medina(reuters.com ↗)
    discuss
  9. A 32-Year-Old Bug Walks into a Telnet Server(watchtowr.com ↗)
    discuss
  10. Republicans Aim to Shield Fossil Fuel Companies from Climate Lawsuits(insideclimatenews.org ↗)
    discuss
  11. Chinese Solar Panels Drop to 12 Cents a Watt, Rooftop Installs Surge Worldwide(oilprice.com ↗)
    1comments
  12. Steam Frame Review – PC Gamer(pcgamer.com ↗)
    discuss
  13. Show HN: Writeably cuts the waffle without losing your meaning or voice(writeably.com ↗)
    discuss
  14. iCloud backup – 5GB available, but I can't use it
    discuss
  15. Show HN: (alter)native-Linux-builder for Nix on Darwin(github.com/quinneden ↗)
    discuss
  16. Show HN: Breakdown: a daily game about guessing data distributions(breakdown.today ↗)
    discuss
  17. Gemini 3.8 Live(ai.google.dev ↗)
    discuss
  18. Software Citations from Replication Files(recite.github.io ↗)
    discuss
  19. One seeded bug, 26 AI agents: all passed the tests, all stayed broken(github.com/vyang472 ↗)
    discuss
  20. Show HN: Whop Video Downloader(chromewebstore.google.com ↗)
    discuss
  21. Show HN: Review agent changes locally before pushing(github.com/marcparadise ↗)
    discuss
  22. Show HN: In-Browser WASM Engine – Analytical Queries in 0.25ms(mlegal-pwa-app.web.app ↗)
    discuss
  23. The Story of Fckgw-RHQQ2-Yxrkt-8TG6W-2B7Q8(twitter.com/davepl1968 ↗)
    discuss
  24. Photos show damage at U.S. positions caused by Iranian missile, drone attacks(cbsnews.com ↗)
    discuss
  25. Single Prompt Knowledge Graphs(corvic.ai ↗)
    discuss
  26. Khipu (Quipu) Field Guide(khipufieldguide.com ↗)
    discuss
  27. Grandma disables a fiber optic drone(twitter.com/united24media ↗)
    1comments
  28. Trump administration releases fifth batch of UFO files(space.news ↗)
    1comments
  29. Show HN: Thurbox – A tmux-based TUI and CLI for local AI agent orchestration(github.com/thurbeen ↗)
    discuss
  30. We cut CDN metadata lookup latency by 91%(vercel.com ↗)
    discuss

Show HN: Review agent changes locally before pushing

2 pointsby 45m agogithub.com
0 comments
I've often found that agents do stupid things when implementing requests. When I only see their changes with limited context in a scroll-by approval blob, it's easy to miss some of that stupidity. In those cases, I'll often only see it when I can review the full change in context in a PR.

This tool adds a local review step, giving the human a final opportunity to be in the loop: your agent commits the changes, pushes to a local 'review' remote, and waits for feedback. The push itself opens a review via post-receive hook, and provides you with a localhost URL to review the changes.

If you request changes via inline comments (like you would for a PR), the agent sees the comments and makes adjustments. It pushes again and the process repeats until you're satisfied.

When you approve the change, your agent can take whatever step is next in your normal workflow - push to open an actual PR, merge locally, etc. These are lightweight instructions (~500 tokens) that should be customized for your workflow.

A quiet thread, for now.Start the conversation on HN ↗