SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

Developer builds crypto wrong-send tool after finding Nansen label quirk saves 99 credits

A developer created a free tool called Sent Wrong to help users determine whether cryptocurrency sent to a wrong address is recoverable. The tool uses Nansen's API to classify addresses and guide users toward the correct next step, such as filing a Binance support ticket or accepting an irreversible loss. During development, the creator discovered that Nansen's cheap profiler endpoints returned wealth tags like 'Token Billionaire' rather than useful exchange entity labels, making them unreliable for the tool's core purpose. The accurate entity label was found instead inside a transaction-lookup endpoint costing just 1 credit, compared to the 100-credit profiler alternative. This finding reshaped the tool's logic from directly querying address identity to strategically selecting transactions most likely to reveal the address's true owner.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer builds token sell planner sized to organic DEX demand, not total volume

A developer created Glidepath, a free open-source tool designed to help non-traders — such as nonprofits receiving crypto donations or freelancers paid in tokens — sell holdings without crashing the market. The tool asks for a token, blockchain, and amount held, then generates a day-by-day selling calendar with each tranche sized to the token's organic buyer demand. The core assumption behind the tool was that professional wallets — Smart Money, whales, funds, and sniper bots — make up a large share of DEX buying volume, making headline volume an unreliable baseline for sizing sells. Using Nansen's wallet-label filters, the developer found that on tested tokens, professional buyers accounted for under 4% of total buys, meaning organic retail demand was far larger than expected. The finding reshaped the product's design, shifting its real value toward calendar-based pacing rather than pro-exclusion filtering.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Why Calibrated Probability Models Outperform Pure Accuracy in Decisions

A blog post by Kartik Pansuriya explores the trade-off between model accuracy and calibration in machine learning decision-making. The piece introduces concepts around JEV (Joint Expected Value) and System One models to argue that well-calibrated probability estimates are more valuable than raw accuracy metrics. Calibration ensures that a model's predicted probabilities reflect real-world likelihoods, which is critical for making reliable decisions under uncertainty. The article contends that optimizing solely for accuracy can be misleading, particularly in high-stakes applications. The post has garnered modest attention on Hacker News, sparking early discussion among practitioners.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Agentic AI System Built to Investigate Fraud End-to-End Using Graph Database

Developers have built an AI-powered fraud investigation agent for the TigerGraph Agentic Fraud Investigation Hackathon, using the HHGOA_IEEE dataset. The agent takes a single fraud alert and autonomously gathers evidence by querying transaction histories, mapping shared devices and regions, and comparing against past cases. It separates reasoning tasks handled by a large language model from rule enforcement, which is managed by deterministic policy code to reduce errors. The system uses vector search and graph traversal to identify fraud rings and retrieve similar historical cases, writing its findings back into the graph so future investigations can learn from them. Each investigation is fully logged and replayable, producing one structured case file per alert with a complete evidence and reasoning trail.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why rerolling AI video prompts fails — and how to fix them with better writing

A developer tested an AI video generation prompt ten times and found the repeated failures stemmed from poor prompt structure, not random model variance. The original one-line prompt asked for both a barista bumping an elbow and a coffee spill, giving the model two competing actions to resolve in just five seconds. Analysis revealed two core mistakes: cramming multiple actions into a single clip, and placing the most important action last, where models tend to underweight it. Rewriting the prompt to focus on a single action — the coffee splash — stated first and framed with specific lens and lighting details, produced the desired result in all ten subsequent generations. The author distilled these findings into a practical checklist and an open-source tool called ai-film-crew, which reviews prompts across seven filmmaking roles before generation.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Four Silent Ways Contact Forms Fail to Deliver Your Leads

Contact forms can appear fully functional while quietly discarding every submission they receive. A common failure occurs when the server's mail() function returns false without alerting the user, making the form seem successful even when nothing was sent. Even when emails do send, they may land in unchecked inboxes or spam folders tied to unmonitored domains. Client-side validation and third-party plugins add further points of failure that never surface during standard design or usability reviews. The core issue is that a form's interface can signal success while the underlying delivery process fails entirely.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How One Developer Built a Battery-Friendly Background Sound Manager for Android

A developer created an Android app called Muffle to automate phone sound profiles based on context, such as prayer times and location, after experiencing the anxiety of potential disruptions in quiet spaces. The app addresses a gap in Android's native Do Not Disturb features, which the developer found too manual, binary, or lacking granular triggers. The core engineering challenge was building a reliable background system that could survive Android's aggressive battery optimizations, including Doze Mode. The solution combined a Foreground Service for time-sensitive sound changes with WorkManager for deferred tasks, keeping the persistent service active only when a high-precision trigger was imminent. Using WorkManager's enqueueUniqueWork with an ExistingWorkPolicy.REPLACE flag ensured task state survived app kills and device reboots, resulting in near-invisible battery usage.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

CVSS 10.0 Flaw in VeloCloud Orchestrator Actively Exploited, Patches Available

A critical vulnerability, CVE-2026-93952, with a maximum CVSS score of 10.0 has been confirmed actively exploited in on-premises VeloCloud Orchestrator (VCO) by Arista Networks. The flaw stems from improper input validation in the VCO web interface, allowing attackers to gain access without operator credentials under certain network conditions. Affected versions include builds 5.2.3.15 and earlier, 6.1.3.7 and below, 6.4.2.7 and below, and 7.0.0.2 and below, while fixed builds 5.2.3.16 and 6.4.2.8 are now available. Cloud-hosted VCO instances were automatically patched and are not affected, but on-premises deployments managing branch SD-WAN infrastructure remain at risk. Security teams are advised to patch immediately, monitor for anomalous files such as vcnode.js, rotate exposed credentials, and establish clear incident response procedures including isolation authority and rebuild plans.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Dev team found AI rules file pointed to deleted script for 24 days, undetected by shape tests

A software development team discovered that one of their AI agent skill files referenced a deleted command for 24 days, during which 28 commit-time shape tests passed without flagging the issue across 142 commits. The problem came to light after the team audited 707 script references across their CLAUDE.md, rules files, and skills, finding that shape tests verify file structure but cannot detect whether named commands still exist in the codebase. To address this gap, the team built a 26-line bash script called check-script-refs.sh that cross-checks all pnpm, npm run, and yarn references in Markdown documentation against the actual scripts defined in package.json. The script identifies missing or stale command references by file and line number, exits with an error code when mismatches are found, and can be integrated into pre-commit hooks or CI pipelines. The audit and tooling were run on September 22, 2026, using Claude Code 2.1.278, and currently report zero missing script references in the team's repository.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

OmniGIF Removes GIF Backgrounds Entirely in the Browser Without Uploading Files

OmniGIF is a browser-based tool that removes backgrounds from animated GIFs without uploading files to any server. All processing — including decoding, color keying, edge cleanup, and re-encoding — runs locally using Canvas ImageData and a GIF encoder Worker. The tool uses chroma-style removal based on RGB color matching within a user-defined tolerance, rather than neural or semantic segmentation, making it fast and deterministic. It handles GIF-specific challenges such as frame-by-frame consistency, a 256-color-per-frame limit, and subjects that share background hues. Users can set the target background color via auto-detection, manual pixel picking, or built-in presets for common cases like white, black, and green screen backgrounds.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Harness Engineering Emerges as Top AI Skill of 2026, Driven by Community Signals

A Thai tech writer known as Nokka analyzed ten widely shared posts on X within a single week in September 2026, finding that contributors across tools, education, and major AI labs were independently converging on the same idea. The consensus: improving AI models alone is insufficient without well-designed surrounding systems, a discipline now called harness engineering. A free GitHub course titled Learn Harness Engineering, launched in late March 2026, has accumulated over 15,800 stars in under six months and covers 14 lessons across 8 hands-on projects translated into 15 languages. The course examines real-world agent architectures from products like Anthropic's Claude Code, OpenAI's Codex, and Inflection's Pi, focusing on five subsystems: instructions, tools, environment, state management, and feedback loops. Supporting signals came from Google, Nous Research, TypeSafe's Jev project, and independent developers, all emphasizing full-system design over prompt writing alone.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Beginner Developer Completes HTML Basics and Begins Learning JavaScript

A beginner web developer documented their eighth day of learning, covering key HTML media elements including iframes and video embedding. After wrapping up foundational HTML and CSS concepts, they transitioned into JavaScript for the first time. The learner explored core JavaScript data types, starting with strings as a basic example. The milestone marks a shift from static webpage structure to learning how to make webpages interactive and dynamic.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer patches wallet-drain vulnerability in AI-powered Solana trading tool BagOS

A security review revealed a two-part vulnerability in BagOS, an MCP server that lets AI agents interact with the Solana-based token launchpad Bags, which could allow a malicious repository to drain a user's crypto wallet. The attack combined a flaw where the server trusted a .env file from the working directory with a login tool that blindly signed any data it received, including full blockchain transactions. Despite 100% test coverage in CI, the flaw went undetected because coverage tracks which lines run, not which inputs are assumed safe. The developer privately drafted a security advisory, patched the bugs on a private fork, and released version 3.0.0 as a breaking change that enforces absolute config file paths and restricts the login tool to signing only Bags' exact sign-in text. All prior versions from 1.0.0 to 2.6.0 have been deprecated on npm, and the advisory has been published as GHSA-g679-3wq7-mh3m.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer builds PowerShell wrapper to run AI scraper safely overnight without supervision

A developer used a PowerShell script — generated by Claude — to automate overnight scraping of ten Reddit communities focused on AI and automation topics. The script managed state, enforced a hard spending cap by querying the provider's actual account spend, and batched subreddit requests in groups of three to avoid Reddit's 403 rate-limiting blocks. Failed or incomplete scrapes were automatically queued for retry without requiring manual intervention, allowing the pipeline to recover missed records while sleeping. Success was measured by the provider's delivered-record count rather than local database row changes, preventing false negatives caused by deduplication. The run expanded the research corpus from roughly 1,000 to 5,846 chunks at a cost of about $34, well below the estimated $118.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Fetching Figma Designs the 'Smart' Way Can Cost Twice as Much, Tests Show

A developer tested Figma-to-code workflows and found that the widely recommended 'outline-first' fetch approach is often more expensive than fetching entire designs at once. On files built without reusable components, the outline-first method cost up to twice as much in API usage compared to a single full fetch. Beyond token costs, the bigger constraint is Figma's hard API quota, which can stretch a 50-screen project from half a day to years on a Starter plan if fetch patterns are inefficient. An additional hidden cost comes from AI agents re-reading all previously fetched data on every turn, compounding expenses over time. The author concludes that optimizing the Figma file structure first and fetching each screen exactly once is the most cost-effective approach.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds Swift Bioinformatics Library BioSwift, Shares Early Progress

A developer has begun building BioSwift, an open-source bioinformatics library for Swift inspired by tools like BioPython and Rust-Bio. The project aims to support DNA and RNA sequence handling, including translation of DNA to RNA and RNA to amino acids. Early development hit a snag when renaming Swift source files in Xcode caused broken references and accidental code deletion, requiring recovery via GitHub Desktop and AI assistance. Despite the setbacks, the developer confirmed that unit tests are passing and core sequence types are functional using Swift's ExpressibleByStringLiteral protocol. The project is in its early stages, with the developer sharing progress through a devlog on DEV Community.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why HTMX May Be the Ideal Framework for AI-Assisted 'Vibe Coding'

A developer analysis published on DEV Community argues that HTMX, a hypermedia-driven library, outperforms React and other SPA frameworks when using AI tools for rapid 'vibe coding' development. The core argument is that modern SPA architectures force AI models to track multiple abstraction layers — from ORM queries to client-side state stores — increasing the likelihood of hallucinations and hard-to-catch bugs. HTMX simplifies this pipeline by having the server return plain HTML directly to the DOM, eliminating JSON serialization layers and client-side state synchronization entirely. Because large language models are extensively trained on server-side templating languages like Django, Rails, and Laravel, they reportedly generate accurate HTMX-compatible code with far fewer errors. The article includes a working Node.js and Express blueprint demonstrating how HTMX attributes can replace hundreds of lines of React state logic in a single compact file.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Independent Audit Flags Medium-High Risk in MEXC Cross-Chain Bridge Worth $5.5B

An independent DeFi security researcher published a risk assessment on September 25, 2026, evaluating MEXC's cross-chain bridge, which holds approximately $5.5 billion in total value locked across Ethereum, multiple Layer-2 networks, and non-EVM chains. The bridge received a composite risk score of 6.2 out of 10, with the lowest ratings in oracle and relayer security and governance upgradeability. Auditors identified a critical re-entrancy vulnerability in the BridgeRelayer contract that could allow unlimited minting of wrapped assets, potentially draining the entire TVL. Additional concerns include a small 12-node validator set vulnerable to collusion, single-key relayer signing susceptible to forgery, and an upgradeable proxy controlled by a single multisig address with no timelock. The report calls for immediate remediation of smart-contract bugs and hardening of relayer infrastructure to reduce the overall risk profile.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

LiteLLM Authentication Flaw Let Single-Character Token Unlock AI Gateway Credentials

A critical improper authentication vulnerability (CVE-2026-59822) in LiteLLM, a widely used AI gateway proxy, allowed attackers to bypass security using any bearer token — including a single character like 'x'. The flaw resided in the OAuth2 header handling code, which returned an empty authentication object on validation failure instead of rejecting the request, effectively granting open access. Attackers could exploit this to list and invoke MCP tools, potentially accessing provider API keys, virtual keys, database credentials, and cloud configurations stored centrally in LiteLLM deployments. Two additional vulnerabilities — a command injection flaw (CVE-2026-42271) and a host header bypass (CVE-2026-48710) — could be chained with the authentication bug to achieve fully unauthenticated remote code execution, a chain linked to the Qilin ransomware group. LiteLLM version 1.84.0 addresses CVE-2026-59822, and users are urged to upgrade immediately.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Fix Python PDF Preview Timeouts by Managing Pages, DPI, and Render Budgets

A technical guide published on DEV Community explains how to debug PDF-to-image conversion timeouts in Python by treating the problem as a render-budget issue rather than a simple timeout configuration. Developers are advised to measure page count, pixel dimensions, and per-page rendering time before adjusting any timeout settings. The article notes that file size is a poor proxy for rendering complexity, since a single-page PDF with a large embedded image can be far more demanding than an 80-page text document. A sample Python implementation demonstrates how to cap page counts and separate fast interactive previews from slower asynchronous batch rendering. The guide recommends using two distinct performance budgets — one for the first visible page and another for remaining pages — to keep preview features predictable under varying document loads.

0 comments Read more at DEV Community
← NewerPage 944 of 3187Older →