Other v0.8 things:
- prompt customization (override per-language packs, prepend house rules)
- OWASP-aligned checklist for manual reviewers
- single Go binary on macOS/Linux/Windows.
brew install mshykov/tap/local-review
v0.8 of local-review: privacy-first AI code reviews — claude/codex/gemini in parallel, no SaaS — plus a bench harness so you can measure which model actually works on your kind of code.
https://t.co/GDB6IID4lC
#golang#buildinpublic
Honest disclaimer: the committed leaderboard uses replay fixtures I hand-crafted. The harness + methodology are the contribution; for real LLM numbers, run it live with your own creds. The dataset is 10 cases — directional, not authoritative.
Good fits:
• Pre-commit hooks (gate exits 2 on Critical/Major)
• Privacy-conscious teams running Ollama
• Multi-cloud orgs not locked to one vendor
• Anyone who wants `local-review init` to scaffold a config in 30s
MIT licensed. Repo: https://t.co/jYJpyo4Nw9
⭐ if useful. 6/6
local-review v0.6.0 is out.
One command. Three reviewers. Your machine. Your keys.
$ local-review review
→ runs claude + gemini + codex in parallel on your git diff
→ merges the findings into one report
→ exits 2 on blocking findings (pre-commit gate ready)
🧵 1/6
Honest privacy posture (it's in the README as a matrix):
• Single-LLM at Ollama → fully offline ✅
• Single-LLM cloud → goes to that one provider
• Multi-LLM → fans out to whichever clouds those CLIs use
The binary itself: no telemetry, no auto-update, no signup. Ever.
5/6
Each reviewer gets the same language-aware prompt pack:
Go, TypeScript, Python, Rust.
Findings come back independently, then a fourth pass consolidates
them — dedup, consensus tags, unified severity rules.
You see what 3 different models actually agree on. Useful signal.
4/6
What I was tired of:
❌ One vendor per tool
❌ Code sent to a SaaS to be reviewed
❌ "AI code review" only happens in CI after the push
❌ Heavy runtimes (Docker, Python envs) just to run a reviewer
local-review is the opposite. Single Go binary, ~10 MB.
2/6
🚀 @get_headway is hiring!
We’re a Ukrainian product edtech company looking for product-motivated talent to help us build a world-class application. 🇺🇦
We need
🔹Product Marketing Manager
🔹Creative Marketing Specialist
🔹User Acquisition Specialist
Links to apply in 🧵👇
Наразі я шукаю до себе в Foundation squad ML Engineer’а, який буде розвивати персоналізова��і системи рекомендацій, покращувати онбординг-кейси за допомогою ML та працювати з LLM/embeddings.
За деталями пишіть в особисті
Щиро вдячний за репост 🤙
Apple says:
Following feedback from the European Commission and from developers, in these releases developers can develop and test EU-specific features, such as alternative browser engines, contactless apps, marketplace installations from web browsers, and marketplace apps, from anywhere in the world. Developers of apps that use alternative browser engines can now use WebKit in those same apps.
Take better care of your Mac with the newly updated @cleanmymac☺️ Here's what's new in the app:
✅ declutter duplicate files
✅ find similar photos
✅ monitor your Mac's health
✅ and more updates
Grab this smart Mac care app on Setapp now👉 https://t.co/cx4MKCVoVa
Setapp Mobile beta is finally here, so if you are in the EU, it's time to explore 50+ quality apps for your iPhone⚡️
Try free for 7 days👉https://t.co/7PKnaH2mBH
Вийшло он��влення SpyBuster, яке допоможе ще краще виявляти небезпечний софт та активність на вашому Mac.
Скачати безкоштовно ту��: https://t.co/V8OLTNoCjH
Дуже дякую команді!
Big news! MacPaw will launch @Setapp Mobile Beta in the EU – the first alternative marketplace for iOS with a curated selection of apps 🚀
Read on for some insider secrets, including the first apps on board: https://t.co/b7Tc2hm7JA