15:[[46530448,{"by":"feross","descendants":125,"id":46530448,"kids":[46530950,46540136,46532318,46533028,46542032,46533287,46539990,46535332,46534677,46536413,46540436,46541935,46536387,46532546,46543885,46536070],"score":205,"time":1767810679,"title":"NPM to implement staged publishing after turbulent shift off classic tokens","type":"story","url":"https://socket.dev/blog/npm-to-implement-staged-publishing"}],[46537095,{"by":"pabs3","descendants":77,"id":46537095,"kids":[46537606,46537750,46538799,46540038,46537158,46537542,46542017,46541904,46540399],"score":181,"time":1767845434,"title":"Go.sum is not a lockfile","type":"story","url":"https://words.filippo.io/gosum/"}],[46515696,{"by":"tbassetto","descendants":1353,"id":46515696,"kids":[46516157,46518038,46520951,46515959,46520115,46519702,46523765,46516135,46529000,46519096,46516096,46518764,46522528,46515862,46524583,46515912,46519436,46516110,46520126,46519591,46515882,46526116,46519792,46523875,46520886,46516335,46519800,46637906,46516198,46521174,46520711,46519474,46516883,46530881,46527561,46531505,46520684,46516076,46519838,46519070,46522517,46527620,46520079,46520800,46533533,46531604,46516316,46525860,46522374,46526755,46524740,46522551,46522174,46527368,46525635,46516495,46560565,46524956,46519888,46524814,46519238,46520619,46516364,46522397,46521337,46528899,46516156,46516111,46525446,46527851,46524235,46519367,46523717,46520633,46520878,46517526,46515961,46531212,46522866,46522105,46524194,46516274,46526764,46516247,46523669,46526401,46527214,46524919,46531598,46516124,46556315,46524630,46521670,46521259,46516040,46521540,46520297,46540223,46520890,46519470,46525437,46522620,46531202,46516336,46521132,46523436,46522560,46529935,46527006,46516497,46525474,46525012,46526803,46516092,46529088,46522643,46522187,46519716,46526628,46527362,46520864,46525600,46525021,46523204,46515820,46520766,46523162,46525127,46515830,46521606,46516021],"score":879,"time":1767721537,"title":"Opus 4.5 is not the normal AI agent experience that I have had thus far","type":"story","url":"https://burkeholland.github.io/posts/opus-4-5-change-everything/"}],[46511578,{"by":"tbrannan","descendants":32,"id":46511578,"kids":[46615317,46513101,46533626,46538054,46512768,46512458,46522425,46513629,46513046,46516017,46512276,46512240],"score":55,"text":"I built DDL to Data after repeatedly pushing back on "just use production data and mask it" requests. Teams needed populated databases for testing, but pulling prod meant security reviews, PII scrubbing, and DevOps tickets. Hand-written seed scripts were the alternative slow, fragile, and out of sync the moment schemas changed.

Paste your CREATE TABLE statements, get realistic test data back. It parses your schema, preserves foreign key relationships, and generates data that looks real, emails look like emails, timestamps are reasonable, uniqueness constraints are honored.

No setup, no config. Works with PostgreSQL and MySQL.

Would love feedback from anyone who deals with test data or staging environments. What's missing?","time":1767703643,"title":"Show HN: DDL to Data – Generate realistic test data from SQL schemas","type":"story"}],[46509688,{"by":"sohanmanju","descendants":3,"id":46509688,"kids":[46509814,46513577],"score":4,"text":"$16","time":1767685148,"title":"Show HN: Yeet – CLI tool to find and reclaim disk space","type":"story","url":"https://github.com/sohanmanju/yeet"}],[46495539,{"by":"shuaimu","descendants":58,"id":46495539,"kids":[46496513,46495963,46495971,46496232,46504616,46496553,46500507,46496576,46496943,46496542,46496043,46500632,46497602,46499811,46495940,46495939,46496334],"score":92,"time":1767589919,"title":"Building a Rust-style static analyzer for C++ with AI","type":"story","url":"http://mpaxos.com/blog/rusty-cpp.html"}],[46496103,{"by":"viveknathani_","descendants":192,"id":46496103,"kids":[46496573,46497305,46497115,46496330,46496488,46497686,46503181,46501895,46506055,46526834,46500357,46497634,46498512,46496833,46499360,46502851,46497742,46498341,46510721,46498939,46496867,46497278,46501479,46498331,46515703,46506904,46509460,46503863,46498637,46497682,46499898,46499720,46498813,46502986,46497202,46500178],"score":717,"time":1767597270,"title":"Databases in 2025: A Year in Review","type":"story","url":"https://www.cs.cmu.edu/~pavlo/blog/2026/01/2025-databases-retrospective.html"}],[46485572,{"by":"indykish","descendants":1,"id":46485572,"kids":[46489028],"score":2,"text":"CacheKit began as an internal feature tightly coupled to Redis and our Rust API framework. We later extracted it into a standalone async crate with a smaller, more explicit public API, while keeping the production constraints that originally shaped it.

I wrote about the process of shipping the crate, including how multiple coding agents were used, where they helped, and where human judgment was still required—especially around correctness, API clarity, and DX.

Write-up:\nhttps://medium.com/@indykish/shipping-a-production-grade-rus...","time":1767508818,"title":"Show HN: CacheKit – an async Rust cache abstraction from a production system","type":"story","url":"https://cachekit.org/"}],[46478932,{"by":"matteo1782","descendants":2,"id":46478932,"kids":[46489003],"score":2,"text":"Phantom Guard is a CLI tool that catches "slopsquatting" attacks before they compromise your supply chain.

The attack vector: AI assistants hallucinate package names → attackers register those names with malware → developers install malware thinking it's legit.

How it works:\n1. Checks if packages exist on registries\n2. Matches against 10+ AI hallucination patterns\n3. Detects typosquats of top 3000 packages\n4. Analyzes metadata (age, downloads, maintainers)

```\npip install phantom-guard\nphantom-guard validate flask-gpt-helper\n# HIGH_RISK: Package not found, matches pattern\n```

Performance: <10ms cached, <200ms uncached.

Try the live demo: https://matte1782.github.io/phantom_guard/

GitHub: https://github.com/matte1782/phantom_guard","time":1767459804,"title":"Show HN: Phantom Guard – Detect AI-hallucinated package attacks","type":"story","url":"https://github.com/matte1782/phantom_guard"}],[46474634,{"by":"akshayaggarwal","descendants":1,"id":46474634,"kids":[46478455],"score":4,"text":"$17","time":1767433306,"title":"Show HN: Boxed – Sovereign exec engine for AI agents (Vercel Sandbox inspired)","type":"story","url":"https://github.com/akshayaggarwal99/boxed"}],[46471246,{"by":"sr-white","descendants":4,"id":46471246,"kids":[46478420,46471409],"score":6,"text":"$18","time":1767399750,"title":"Show HN: Snowflake Emulator – Local Snowflake Development with Go and DuckDB","type":"story","url":"https://github.com/nnnkkk7/snowflake-emulator"}],[46477561,{"by":"jaz1","descendants":2,"id":46477561,"kids":[46478374],"score":4,"text":"$19","time":1767453155,"title":"Show HN: Jetbase – A Python database migration tool (Alembic alternative)","type":"story","url":"https://github.com/jetbase-hq/jetbase"}],[46456682,{"by":"pavel_lishin","descendants":34,"id":46456682,"kids":[46460306,46458739,46457642,46460269,46461570,46457460,46457795,46457877,46458237,46457902],"score":75,"time":1767292465,"title":"Building an internal agent: Code-driven vs. LLM-driven workflows","type":"story","url":"https://lethain.com/agents-coordinators/"}],[46456452,{"by":"EddieDean","descendants":3,"id":46456452,"kids":[46456459,46457846],"score":3,"time":1767291211,"title":"Show HN: Arrayops – 100x faster Python array operations with Rust","type":"story","url":"https://github.com/eddiethedean/arrayops"}],[46446857,{"by":"mehdig10","descendants":4,"id":46446857,"kids":[46448592,46447788],"score":1,"text":"Hi HN! I built Basehook to automate common webhook patterns I've had to manually implement across multiple projects: buffering, smart deduplication, and payload inspection.

The core idea: webhooks are stored in Postgres, grouped by thread ID and revision number. You can then consume them either:

- One by one (process every update in order)

- Buffered (skip to the latest revision, ignore outdated ones)

This is useful when you only care about the final state. Example: A Slack bot receives 5 edits to the same thread. Instead of processing all 5, you can only process the thread once and call the Slack API to fetch the full thread content once.

It works by extracting thread IDs and revision numbers from webhook payloads via configurable JSON paths, so it supports any format (Slack, GitHub, Shopify, or custom).

Built with FastAPI, React, and PostgreSQL. The UI lets you configure webhooks and view all received payloads, with the ability to re-queue if needed.","time":1767205907,"title":"Show HN: Basehook – Webhook management system built on Postgres","type":"story","url":"https://github.com/mehdigmira/basehook"}],[46444159,{"by":"prxtenses","descendants":2,"id":46444159,"kids":[46447752],"score":1,"text":"$1a","time":1767190137,"title":"Show HN: Repair-JSON-stream – Fix broken JSON from LLM streaming (1.7x faster)","type":"story","url":"https://github.com/prxtenses/repair-json-stream"}],[46435308,{"by":"keepamovin","descendants":221,"id":46435308,"kids":[46436937,46438109,46440119,46437110,46436853,46436422,46439282,46438955,46443012,46436543,46442123,46436488,46436985,46436803,46438509,46439356,46437026,46487080,46443346,46438681,46436571,46440398,46443496,46440374,46438345,46446367,46442902,46438637,46440946,46443965,46439422,46437009,46439002,46438133,46439263,46442280,46435316,46436910,46435771,46438884],"score":730,"text":"Community, All the HN belong to you. This is an archive of hacker news that fits in your browser. When I made HN Made of Primes I realized I could probably do this offline sqlite/wasm thing with the whole GBs of archive. The whole dataset. So I tried it, and this is it. Have Hacker News on your device.

Go to this repo (https://github.com/DOSAYGO-STUDIO/HackerBook): you can download it. Big Query -> ETL -> npx serve docs - that's it. 20 years of HN arguments and beauty, can be yours forever. So they'll never die. Ever. It's the unkillable static archive of HN and it's your hands. That's my Year End gift to you all. Thank you for a wonderful year, have happy and wonderful 2026. make something of it.","time":1767114119,"title":"Show HN: 22 GB of Hacker News in SQLite","type":"story","url":"https://hackerbook.dosaygo.com"}],[46379765,{"by":"flyaway123","descendants":25,"id":46379765,"kids":[46434657,46435736,46439970,46436950,46435681,46438091,46435769],"score":143,"time":1766613897,"title":"Postgres extension complements pgvector for performance and scale","type":"story","url":"https://github.com/timescale/pgvectorscale"}],[46415045,{"by":"ubermenchh","descendants":4,"id":46415045,"kids":[46419828,46425840],"score":5,"text":"I built this to understand how vLLM works internally.","time":1766959981,"title":"Show HN: Mini-vLLM in ~500 lines of Python","type":"story","url":"https://github.com/ubermenchh/mini-vllm"}],[46423017,{"by":"saisrirampur","descendants":2,"id":46423017,"kids":[46425808],"score":2,"time":1767029557,"title":"Postgres and ClickHouse forming the default data stack for AI","type":"story","url":"https://thenewstack.io/postgres-clickhouse-the-oss-stack-to-handle-agentic-ai-scale/"}],[46410656,{"by":"rostislav_dugin","descendants":7,"id":46410656,"kids":[46411646,46413218,46414341],"score":10,"text":"I had a post in the start of December about Postgresus 2.0: self hosted tool for PostgreSQL. Since then the project changed the name and added support of other databases: MySQL, MariaDB and MongoDB.

Website - https://databasus.com

GitHub - https://github.com/databasus/databasus

Main features:

- Scheduled backups for multiple PostgreSQL, MySQL, MariaDB and MongoDB databases

- Storage to S3, Cloudflare R2, Google Drive, Azure Blob, NAS, SFTP, rclone, etc.

- Notifications about backup status via email, Telegram, Slack, Discord, MS Teams and customizable webhooks

- Works with both self-hosted and managed PostgreSQL (RDS, Cloud SQL, etc.)

- Runs as a single Docker container or via Helm on Kubernetes, can be installed via script

- There are workspaces and role-based access with audit logs","time":1766925538,"title":"Show HN: Databasus – open-source backup tool for PostgreSQL, MySQL and MongoDB","type":"story","url":"https://databasus.com"}],[46412348,{"by":"Arulnidhi_k","descendants":4,"id":46412348,"kids":[46426854,46414311],"score":1,"text":"Aegis Memory is an open-source, self-hostable memory layer for multi-agent AI systems.

v1.2 adds Smart Memory - a two-stage pipeline that automatically decides what's worth storing:\n1. Fast rule-based filter catches obvious noise (greetings, "thanks", etc.)\n2. LLM extracts atomic facts only when the filter passes

This saves ~70% of extraction costs while keeping memory high-quality.

Try it in 15 seconds:\n pip install aegis-memory\n aegis demo

GitHub: https://github.com/quantifylabs/aegis-memory

Happy to answer questions about multi-agent memory architecture.","time":1766940310,"title":"Show HN: Aegis Memory v1.2 – We solved \"what's worth remembering\" for AI agents","type":"story","url":"https://github.com/quantifylabs/aegis-memory"}],[46402797,{"by":"dkalola","descendants":2,"id":46402797,"kids":[46405021],"score":1,"text":"Just a quick experiment I thought of building. Limited practical usage, for sure.

A Python library that uses LLMs as comparators for semantic sorting. Sort any list by meaning, tone, complexity, urgency, or any criteria expressible in natural language. This library uses an OpenRouter API key to interface with LLM providers.","time":1766852085,"title":"Show HN: LLM Sorter – Python package to sort lists of items using LLM calls","type":"story","url":"https://github.com/kalolad1/llm-sorter"}],[46400118,{"by":"cozymori","descendants":2,"id":46400118,"kids":[46404988],"score":4,"text":"$1b","time":1766823036,"title":"Show HN: VectorWave – Turn Python function execution into searchable vector data","type":"story","url":"https://github.com/Cozymori/VectorWave"}],[46404312,{"by":"abdulbasitali","descendants":3,"id":46404312,"kids":[46404935],"score":3,"text":"$1c","time":1766862982,"title":"Show HN: AgentFuse – A local circuit breaker to prevent $500 OpenAI bills","type":"story","url":"https://github.com/AbdulBasitA/agent-fuse"}],[46393992,{"by":"zdw","descendants":459,"id":46393992,"kids":[46397584,46395821,46396087,46396092,46396652,46395919,46397254,46396250,46396483,46395819,46400086,46395737,46396962,46396537,46397451,46397383,46398449,46405150,46395881,46396957,46396210,46397428,46399435,46396842,46397356,46396725,46402940,46403576,46396124,46401738,46411261,46397950,46396462,46399197,46400850,46396814,46405557,46408801,46397433,46410026,46396234,46398712,46396188,46402112,46398876,46399097,46398882,46414345,46396160,46395971,46400672,46396079,46397085,46399622,46399358,46397163,46396767,46400205,46396169],"score":1290,"time":1766769187,"title":"How uv got so fast","type":"story","url":"https://nesbitt.io/2025/12/26/how-uv-got-so-fast.html"}],[46396267,{"by":"simonebrunozzi","descendants":2,"id":46396267,"kids":[46396530,46398972],"score":11,"time":1766783196,"title":"Uv: An Fast Python Package Manager","type":"story","url":"https://www.janestreet.com/tech-talks/uv-an-extremely-fast-python-package-manager/"}],[46398092,{"by":"saisrirampur","descendants":2,"id":46398092,"kids":[46398965],"score":1,"time":1766797056,"title":"Postgres and ClickHouse forming the default data stack for AI","type":"story","url":"https://thenewstack.io/postgres-clickhouse-the-oss-stack-to-handle-agentic-ai-scale/"}],[46398904,{"by":"tanelpoder","descendants":1,"id":46398904,"kids":[46398959],"score":2,"time":1766807097,"title":"Pyleak: Detect asyncio event loop blocking with stack traces in Python","type":"story","url":"https://github.com/deepankarm/pyleak"}],[46397566,{"by":"maheshvaikri99","descendants":14,"id":46397566,"kids":[46408116,46450002,46397595,46460740,46398950,46397785,46398303],"score":7,"text":"$1d","time":1766792308,"title":"Show HN: ISON – Data format that uses 30-70% fewer tokens than JSON for LLMs","type":"story","url":"https://github.com/maheshvaikri-code/ison"}]]

user:	dmarwicke
created:	December 27, 2025
karma:	6