Programming

ProgrammingDEV Community ·

BeeLadybug Ironic: A Browser Tool Helps Developers Decompress After Tough Coding Days

A developer has built an experimental browser-based tool called BeeLadybug Ironic, designed to help programmers mentally unwind after stressful coding sessions. Inspired by the rubber-duck debugging technique, the project extends the concept of self-inspection from code to personal frustration. Users are greeted by a bee-ladybug mascot and offered simple prompts to reflect on whether they are genuinely stuck or simply burnt out. They can type out their frustrations and watch the text dissolve into pixels as a symbolic release ritual. The tool runs entirely in the browser's RAM with no accounts, servers, or external tracking, and is currently in the prototype stage seeking community feedback.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Free Browser Tool Lets You Check If Your VPN Is Leaking Your Real IP

A developer has built a free, browser-based tool called EvoTechTool that tests whether a VPN is leaking a user's real IP address via WebRTC, a technology that powers video calls and peer-to-peer communication. WebRTC can expose a device's true public IP even when a VPN is active, because it fetches network addresses through a separate channel that bypasses the VPN tunnel. The tool uses the browser's RTCPeerConnection API to detect IP addresses disclosed during WebRTC's connection process and compares them against the IP the VPN reports. All processing happens locally in the browser with no data sent to external servers, according to the developer. Beyond WebRTC, the tool also flags other common privacy risks such as DNS leaks, IPv6 exposure, browser fingerprinting, and time zone or language data that can reveal a user's location.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

ephemora-cell 1.0.4 Launches on PyPI as a Stateless MCP Tool Server

ephemora-cell version 1.0.4, a stateless Model Context Protocol (MCP) tool server, is now available on PyPI. The release eliminates session state and initialization handshakes, allowing servers to be restarted or load-balanced between calls without clients noticing. A key bug fix ensures the get-policy report accurately reflects actual filesystem access granted during execution, closing a gap where sandbox directories were previously underreported. The update also resolves a version mismatch issue by single-sourcing the version number, ensuring consistency between the package metadata and runtime. Backward compatibility is maintained for handshake-era clients such as Claude Desktop, VS Code, and Codex.

0 comments Read more at DEV Community

ProgrammingHacker News ·

FAA Halts East Coast Flights After Fiber Line Cut Disrupts Communications

The FAA temporarily halted flights at busy East Coast airports due to communication issues, according to a Reuters report dated September 21, 2026. Authorities attributed the disruption to a severed fiber line affecting air traffic communications. The ground stop impacted operations at multiple airports along the East Coast. No further details about the scope, duration, or responsible party were available from the provided source.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

VS Code AI Assistants in 2026: Free Tier Limits, Top Tools, and Privacy Risks

By mid-2026, AI assistants have become the most-installed category of VS Code extensions, with 41% of users activating them weekly, according to GitHub data. Tools like GitHub Copilot, Codeium, and Amazon Q dominate the market, though all free plans impose daily usage caps that most developers exhaust within two hours of focused work. Codeium has seen rapid growth, reaching 3.5 million active installs — a 190% rise in 12 months — while outperforming Copilot on Python test generation in independent benchmarks. A key privacy concern has emerged: most free-tier assistants transmit code snippets to remote servers by default, posing potential risks for developers working on proprietary codebases. Tabnine remains the primary option for fully local inference, though its free plan excludes chat and advanced features.

0 comments Read more at DEV Community

ProgrammingHacker News ·

404 Media Seeks Public Help to Map Police Surveillance Camera Use Across US

Investigative outlet 404 Media is crowdsourcing information from the public to document how local police departments are deploying surveillance technology. The effort focuses specifically on Axon license plate recognition cameras, which are increasingly being adopted by law enforcement agencies. Residents are being asked to share what they know about camera installations in their cities or towns. The initiative aims to build a broader picture of how widespread such surveillance infrastructure has become across the United States.

0 comments Read more at Hacker News

ProgrammingHacker News ·

Suspicious NPM Math Library Found Hiding Encrypted Loader in Package

Security researchers at SafeDep discovered an NPM package posing as a math library that conceals an encrypted loader within its code. The package raises red flags because legitimate math libraries have no justifiable reason to include obfuscated or encrypted execution components. This technique is commonly associated with malicious packages that attempt to evade static analysis and security scanning tools. The finding highlights ongoing supply chain security risks within the NPM ecosystem, where threat actors disguise malware as innocuous utility libraries. Developers are advised to scrutinize dependencies carefully, especially those with unusual or unexplained code components.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

How vLLM Solves GPU Memory Bottlenecks in Large Language Model Inference

Modern LLM text generation is primarily a memory bandwidth problem, not a compute problem, with GPUs often running at low utilization while shuffling model weights and KV cache data during the decode phase. Standard PyTorch inference pipelines suffer from memory fragmentation, static batching inefficiencies, and high driver overhead when handling multiple concurrent users. vLLM addresses these limitations through two core innovations: PagedAttention, which manages KV cache memory more efficiently, and continuous iteration-level batching, which schedules new requests as soon as any sequence completes. Benchmarks for the article were conducted on dedicated NVIDIA H100 and H200 GPU clusters, using models such as Qwen3.6-27B as practical examples. The piece provides a technical deep dive into why naive inference stacks fail at scale and how vLLM's operating-systems-inspired engineering overcomes those constraints.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Google, OpenAI and X Move Ads Into AI Chat Interfaces Across Platforms

Major tech platforms including Google, OpenAI, and X are testing advertising formats embedded directly within conversational AI interfaces rather than in traditional placements. Google announced Gemini-powered ad formats at its Marketing Live 2026 event, designed to appear within AI-driven search and discovery experiences. OpenAI has made ChatGPT Ads available in several regions, placing ads within and alongside chat threads based on conversation topics and user interactions. X has separately experimented with contextual ad units integrated into social post conversations and discussion flows. The shared trend signals a cross-platform shift toward dialogue-aware advertising, where relevance depends on matching a user's conversational intent rather than interrupting their attention.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Dart vs Freezed: Benchmark Tests Four Ways to Write a Data Class

A developer benchmarked four approaches to writing a data class in Dart — Freezed 4, Equatable, pure Dart 3.13, and Freezed 4 with primary constructors — using a shared suite of 16 tests on a single Note model. The comparison measured lines written, lines generated, build time, and how many places need updating when a new field is added. Dart 3.13, which made primary constructors stable on 12 August 2025, has been gradually closing the gaps that originally made Freezed necessary back in 2020. No single approach outperformed the others across all criteria, with each breaking down in different scenarios. The findings point toward a recommended default approach, with identified exceptions where that default is not the best fit.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Domain Claim Protocol Aims to Define Clear Boundaries for Autonomous AI Agents

As software systems increasingly rely on autonomous agents that dynamically select tools and services, a key engineering challenge has emerged: defining what each component is actually permitted to do. Unfurl Systems Lab has published an open draft specification called the Domain Claim Protocol (DCP), currently at version 0.2.0-draft, to address this gap. DCP allows each component to declare its role, capabilities, dependencies, explicit refusals, and fault behaviors in a structured, machine-readable format. Unlike API schemas or prompt instructions alone, a DCP claim is designed to be inspectable by compilers, runtimes, and human reviewers before components are assembled into a workflow. The protocol aims to replace implicit, drifting boundaries in complex agent systems with enforceable, transparent contracts that clarify ownership of business decisions and failure handling.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Kern Promises Local Code Intelligence to Cut AI Token Costs by Up to 80%

A new open-source tool called kern, built in pure Go, aims to reduce the token overhead incurred when using AI coding assistants such as Claude Code, Cursor, and Codex. Rather than feeding entire codebases into large language models, kern uses AST-based parsing to deliver surgical context slices estimated at 500–2,500 tokens per query, compared to much larger conventional approaches. The tool operates as a single self-contained binary with local-first, zero-telemetry architecture that masks PII and credentials before any data reaches an LLM. A comparison with rival tools Graphify and Code Review Graph positions kern as better suited for active daily development, while the others remain preferable for multimodal documentation and CI/PR risk scoring respectively. Kern is available on GitHub under the handle JayveerPrajapati/kern and supports installation on macOS, Linux, and Windows.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How to Prevent AI Agents from Gaming CI Test Scores with a Locked Contract File

Software teams using AI coding agents face a subtle risk: a passing CI build does not confirm genuine code quality if the agent can alter tests, fixtures, or runner configuration to manufacture a green result. A proposed mitigation involves recording a score contract from the parent commit — capturing fixture hashes, runner config hashes, assertion floors, test-collection floors, and time-bounded flake leases — before the agent's branch is evaluated. The contract file and scorer tooling are then placed outside the agent's write set, so any patch touching those paths is automatically rejected. Content hashing of fixtures and pytest configuration files is recommended over path-based denylists, which can be bypassed by adding sibling files. The approach acknowledges a known limitation: if the parent test suite is already weak or tautological, the contract will preserve that weakness rather than correct it.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Four Evidence Signals Edtech Teams Need for Reliable API Key Access Reviews

A technical guide outlines a four-signal framework for accurately attributing API keys to their owning services during edtech access reviews. The method combines a stable key identifier, verified caller identity, workload deployment records, and per-key request events to build a defensible evidence bundle. Reviewers are cautioned against relying on easily spoofed signals like source IP addresses or user-agent strings, which cannot confirm service ownership on their own. The guide highlights a common risk where a single credential mounted across multiple workloads — such as a course-search service and an evaluation job — can silently grant unintended production access. It recommends keeping attribution uncertainty visible in records and rotating or re-scoping shared credentials only after all consumers have been identified, to avoid accidental outages.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer Builds Serverless AI News Bot Using Node.js, Gemini, and GitHub Actions

A developer has created an open-source serverless bot called Chaos Routine Bot that automatically delivers AI-generated technology news briefings each morning. The pipeline scrapes from over a dozen sources — including Hacker News, TechCrunch, and Ars Technica — randomly selecting two per run, then filters articles using a tech-relevance keyword system before processing. Extracted article text is condensed via Mozilla Readability and Compromise NLP, with only the first three sentences passed to Google's Gemini 2.5 Flash model to generate a sarcastic editorial-style summary. If the Gemini API is unavailable, the bot gracefully falls back to offline NLP-generated summaries, ensuring uninterrupted output. The final briefing is distributed simultaneously to Telegram and Discord, while a JSON feed, RSS output, and a Vercel-hosted dashboard are also maintained.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Citrix NetScaler SAML Bypass CVE-2026-19490 Exposes Hundreds of Thousands of Gateways

A critical SAML authentication bypass vulnerability, CVE-2026-19490, has been identified in Citrix NetScaler ADC and Gateway, carrying a CVSS score of 9.8. The flaw allows attackers to forge sessions by presenting fraudulent SAML assertions that the gateway accepts as valid, potentially granting unauthorized access to all applications behind it. An emergency patch was released in August 2026, but active exploitation was still being reported in subsequent weeks. Internet scans conducted via ZoomEye on September 22, 2026, found between approximately 117,000 and 239,000 publicly reachable NetScaler assets, depending on the query fingerprint used. Security teams are advised to verify patch levels on every NetScaler instance handling SAML authentication and to audit logs for anomalous or forged authentication assertions.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer Tests Jev AI Model for Faster Browser and Computer Automation

A developer gained early access to Jev, a system-one AI model from typesafe.ai that is designed around a typed decision interface rather than traditional text generation. Unlike general-purpose LLMs such as GPT or Claude, Jev accepts a structured state and returns typed outputs — choices, scores, or yes/no decisions — making it suited for fast, local decision-making tasks. The developer is experimenting with applying Jev to computer-use scenarios, using either screenshot-based OCR pipelines or OS accessibility APIs to feed structured inputs into the model. Typesafe.ai describes itself primarily as a data research company and claims the model was trained entirely on synthetic data with near-zero hallucination. However, the developer cautions that being type-safe does not inherently prevent hallucinations, and that the confidence scores are derived from classification probabilities and may not suit every use case.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How to Build Reliable Email Audit Trails Without Relying on Webhooks

A software architecture guide recommends placing an outbox layer between application business events and email APIs to ensure durable, auditable delivery records for messages like password resets and order receipts. The approach assigns stable message keys, stores provider-returned message IDs, and uses a polling mechanism to reconcile unresolved delivery states rather than depending on webhook callbacks. Password reset emails and order receipt emails require different retry policies due to their distinct purposes — one involves a time-sensitive security token, the other documents a settled financial transaction. The guide emphasizes that an email provider's acceptance response only confirms receipt of the request, not actual delivery, and all subsequent delivery events should be stored in an append-only history. API selection should therefore prioritize providers that expose queryable, stable message identifiers over those that only support dashboard lookups or push-only event models.

0 comments Read more at DEV Community

ProgrammingHacker News ·

Apple's Abandoned Copland OS Prototype Now Runs in a Web Browser

A developer has managed to get Apple's Copland D11E4 build running directly inside a web browser. Copland was an experimental next-generation operating system that Apple developed in the mid-1990s but ultimately cancelled before release. The project was shared on pagetable.com, a site known for low-level computing and retrocomputing research. This achievement allows anyone with a browser to experience the historic, never-shipped Apple OS without dedicated hardware or emulator software.

0 comments Read more at Hacker News

ProgrammingHacker News ·

Research Addresses Babbling-Idiot Failures in Time-Triggered Communication Systems

A technical paper published on IEEE Xplore examines a critical fault type known as the 'babbling-idiot' failure in time-triggered communication systems. This failure occurs when a node transmits data outside its allotted time slot, potentially disrupting the entire network. The research focuses on strategies to detect and prevent such faults to improve system reliability. The paper was shared on Hacker News, where it attracted modest attention with 5 points and one comment.

0 comments Read more at Hacker News

← NewerPage 1041 of 3196Older →