SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

Agents Don't Need Memory, They Need Documentation: A Practical AGENTS.md Playbook

Mấy tháng gần đây, mình thấy team nào dùng AI coding agent (Claude Code, Codex CLI, Cursor, Aider...) cũng gặp chung một vấn đề. Hôm nay agent làm rất tốt. Sang hôm sau, nó lại quên sạch, dùng sai convention, chạy sai lệnh test, sửa nhầm file generated. Phản xạ đầu tiên của nhiều người là đi tìm giải pháp "memory": vector DB, plugin nhớ hội thoại, tóm tắt session. Nhưng sau khi thử khá nhiều cách, mình đồng ý với một bài đang hot trên Hacker News: agent không cần memory, nó cần documentation.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

I Traced AgentCore's Package Injection Bug Through 2 Fixes

I was updating my agent sandbox audit scripts over the weekend and stumbled into something that changed how I look at every SDK helper that shells out. AWS's Bedrock AgentCore Python SDK, the thing you use to run code in Amazon's Code Interpreter sandboxes, had the same injection bug fixed twice in the same function. Both fixes were bypassable, each in a different way. This matters right now because BeyondTrust's Phantom Labs team published their full research on it yesterday (Sep 28), and as of this morning I can find no other write-up anywhere: zero threads on Hacker News, zero articles on D

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Make SSH Use a Password Instead of a Key

Need to test password-based SSH login, or connect to a host without offering your usual key? You can override the client’s authentication choices for a single connection without changing or deleting any keys. ssh -o PreferredAuthentications=password -o PubkeyAuthentication=no user@host Replace user and host with the account and server you want to reach. These options apply only to this invocation of ssh. PreferredAuthentications=password tells the client to use the SSH password authentication method.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Building a German Speaking Partner with ElevenLabs Agents and Cloudflare Workers

Anyone who has learned a foreign language knows the gap between a passed exam and real fluency with native speakers. Closing it takes practice with someone at native level, and not everyone has a native-speaking partner or the budget and time for endless lessons. So I asked myself: what if an AI voice agent could be that partner? Someone to have casual conversations with, like a buddy who speaks at native level, never loses patience, and points out your mistakes when they matter. I built it with ElevenLabs agents and Cloudflare Workers.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

A browser QR scanner without BarcodeDetector or a CDN

An in-page QR scanner sounds like a thin wrapper around a camera API. The first version often is: request a camera, hand frames to BarcodeDetector, and continue when it returns a QR value. That design has an awkward failure mode. The camera can work perfectly while QR detection is unavailable. A browser may expose getUserMedia() but not BarcodeDetector, so capability detection turns into a message telling the user to leave the app and find another scanner.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Bob Cringely Has Died

I heard from a friend of the family that Bob passed away in his sleep early Saturday. Very sad news. Bob, who's real name was Mark Stevens, was an early employee of Apple and was best known for his PBS documentaries, especially "Triumph of the Nerds". He will be missed. Comments URL: https://news.ycombinator.com/item?id=49949438 Points: 37 # Comments: 3

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

SaaS ไทยในยุค AI: แผน 90 วันสำหรับผู้ประกอบการที่ยังไม่พร้อมเปลี่ยนทั้งระบบ

โดย Nokka (นก-กา) | 3 ตุลาคม 2569 สมมติว่าคุณเป็นเจ้าของธุรกิจขนาดกลางในไทย มีพนักงาน 80 คน ใช้ซอฟต์แวร์อยู่ราว 12 ตัว ทั้งบัญชี เงินเดือน หลังบ้านขาย และระบบดูแลลูกค้า คุณได้ยินมาว่า AI agent กำลังเปลี่ยนทุกอย่าง และคุณไม่รู้ว่าจะเริ่มจากตรงไหน ข่าวดีคือคุณไม่ได้เริ่มจากศูนย์ เพราะตัวเลขล่าสุดบอกว่าธุรกิจไทยจำนวนมากกำลังทำสิ่งเดียวกันอยู่ และผมจะเล่าให้ดูว่าผมอ่านตัวเลขชุดนี้ว่าอย่างไร จากผลสำรวจ UOB Business Outlook Study 2026 (Thailand) H1 ซึ่งสอบถามเจ้าของธุรกิจและผู้บริหารระดับตัดสินใจ 265 ราย พบว่า 71% ของธุรกิจไทยลงมือใช้ AI แล้ว สูงกว่าค่าเฉลี่ยภูมิภาคอาเซียน และ 89% วางแผนเพิ่มงบ AI ใ

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

I Traced vm2's 9.5 Sandbox Escape to a Missing Regex Boundary

allowlisted: .../node_modules/foo/index.js recorded as: ^\.\.\.\/node_modules\/foo `/tmp/vm2-prefix-case/node_modules/${module}` }, context: 'host', // NodeVM default }); // guest code: // require('foo') -> prefix recorded: ^/tmp/vm2-prefix-case/node_modules/foo // require('/tmp/vm2-prefix-case/node_modules/foo2/index.js') -> passes the prefix The first require is legitimate: the resolver answers with /tmp/vm2-prefix-case/node_modules/foo, and vm2 records that prefix. The second require points at foo2, a completely different, non-allowlisted package whose path happens to start with the same ch

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

The transmit chain: the easy half, done right

Part 2 of "Building a MIL-STD-188-110 HF Modem in Modern C++." Start with the overview, or Part 1, the rules I build by. The code for this article is on GitHub: kg7hf/modem-core. At the end of Part 1 I called the transmitter the easy half. Encoding is deterministic, there is no channel to undo, and the standard spells out every step. I still hold it to the same discipline as the receiver.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Browser Use vs Skyvern Self-Hosted: Setup, Trade-offs, and Completion Verification

Why we evaluated Browser Use and Skyvern We did not bring Browser Use and Skyvern into our lab to see whether an agent could click a button in a polished demo. We wanted to answer a more expensive question: can either system complete a recurring login-to-download workflow without creating a permanent human repair queue? Agent-reported done is not verified completion A strict schema receipt still passed for a missing and a corrupted file, so a workflow should count as complete only when page state, business identifier, and artifact size and digest are checked independently of the agent's own su

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Open-Source PDF Editor for React. Here’s How to Add PDF Editing to Your App

Adding a PDF viewer to a React application is relatively easy. Adding an actual PDF editor is a different story. I recently needed a way to let users edit PDF documents directly inside a React application. I wanted something that could be embedded into an existing app without sending users to an external PDF editing service. That led me to build react-web-pdf-editor — a React package for adding browser-based PDF editing capabilities to React applications.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Menu Digitization — Node.js Upload-Time Processing vs On-Demand Image Cleanup

Short answer Short answer: for restaurant menu digitization, inspect metadata at upload, but defer expensive image cleanup until a dish is actually indexed or viewed. That split keeps searchable dish data predictable without making every upload wait on a full media pipeline. It starts with one rule: It failed once. I keep a long-lived rule: metadata is a gate; pixels are a workload. A menu photo with the wrong orientation, an impossible color profile, or a 40 MB payload should be rejected or normalized immediately.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

OpenAI-Compatible Image Generation: How to Route Multiple Providers with Fallbacks

Short answer: put an OpenAI-compatible image request behind a small dispatcher, load image-capable model IDs from a catalog at deploy time, and keep the primary and fallback in configuration. Fail over only on transient failures. A rejected prompt is not a reason to ask another provider. This matters in a fintech hiring system that scores candidates against a job rubric. The scoring result should remain structured data; text-to-image belongs downstream, where it can produce a recruiter-facing report visual without changing the score.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Your LLM Bill Jumped After You Added Context: Find the Cache Miss Before You Downgrade the Model

If your LLM API spend climbed after you added retrieval, a longer system prompt, or tool definitions, the cause is almost always input tokens being reprocessed at full price on every request — not the model you picked. Check the cache fields in the response usage object first: if cached reads are zero across repeated requests that share a prefix, you have a bug, not a pricing problem. Fixing prefix stability and breakpoint placement is free; downgrading the model costs you quality. The first thing to do is stop reasoning about the bill and start logging per-request token accounting. On the Ant

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

I Read Claude's Biology Discovery as a Pipeline: 5 Lessons

Anthropic posted on September 23 that Claude agents found a new enzyme system hiding in phage DNA. The numbers in the post: roughly 950 agents, 21 hours, 210 million tokens, one giant database of DNA sequences. One agent noticed a repeating pattern next to a gene that looked odd. Humans did the lab work, named the system ART, and released a pre-print. When I checked around 1 PM Singapore time the Hacker News thread sat at 567 points and 588 comments, and most of the fight is about one word: discovery.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

The BIN table said debit. The issuer said credit.

Added card-surcharging logic for a US merchant: debit cards skip the 3% fee, credit cards get it, per the usual state rules. Classification ran off a static BIN range table refreshed quarterly from our processor. Surcharge complaints started trickling in — cardholders insisting their debit card got charged a credit surcharge. Pulled the flagged transactions: roughly 1 in 180 had a BIN that our table listed as credit, but the issuer's own authorization response carried a debit indicator in the response fields. Turns out prepaid and co-branded cards get reissued under ranges that shift between d

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

I Audited the MCP SDK OAuth Fix: 3 Checks Upgrading Misses

Last week the official MCP Python SDK shipped a fix for a credential theft bug, and the fix itself has a trap that a version check will not catch. I was researching the disclosure for my own MCP servers and the remediation notes stopped me cold: patch the package, and part of the vulnerability survives unless you also change how you construct your OAuth providers. Here is the short version. On September 28, a GitHub advisory landed on the modelcontextprotocol python-sdk repo: GHSA-qx49-fqc8-xw99, titled "OAuth client could send credentials to an authorization server chosen by the MCP server."

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Minimalist Backend Automation Techniques: An Engineer's Insight

Minimalist Backend Automation Techniques: An Engineer's Insight Here's the brutal truth: most tech solopreneurs eventually drown in a sea of server crashes and API limits, suffocating under the weight of their own systems. While the tech landscape offers vast potential for innovation and efficiency, the road to automation is often cluttered with obstacles that require meticulous navigation. For those navigating these technical waters, the promise of streamlined backend operations offers a beacon of hope. Entrepreneurs in the tech world are all too familiar with the intricacies of backend opera

0 comments Read more at DEV Community
← NewerPage 669 of 3153Older →