SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

Venus Core Pool Governance Rated High Risk With Nine Attack Vectors Identified

A security audit of Venus Core Pool, a DeFi lending protocol with approximately $1.28 billion in total value locked, has identified nine distinct governance attack vectors and assigned an overall risk rating of 7 out of 10. The review, dated September 18, 2026, focused on the on-chain governance flow covering proposal creation, voting, execution, and contract upgrades. Among the most critical findings are a flash-loan-driven voting power inflation attack scored 9/10 and a timelock bypass via re-entrancy scored 8/10, both considered technically feasible. Auditors noted that the governance subsystem had not been thoroughly examined since the v2.3 upgrade in March 2025, leaving key administrative functions exposed. The report attributes the elevated risk to a combination of a large treasury, powerful admin controls, and a relatively permissive quorum model that could be exploited by a coordinated or well-funded adversary.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer builds read-only AI agent that audits AWS accounts in under two minutes

A developer has created an AI agent capable of auditing an AWS account for security and cost issues, identifying problems such as publicly exposed S3 buckets, open SSH ports, disabled GuardDuty, and wasted spend on idle Elastic IPs. The entire audit takes roughly two minutes to complete. The agent is deliberately restricted to a read-only IAM identity, meaning AWS itself blocks any write, delete, or modify actions before they can execute. The build relies on just one JSON configuration file and a Markdown checklist, requiring no Python frameworks or vector databases. The full source code has been published on GitHub for others to deploy against their own AWS accounts.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Study: Human Brain May Have Evolved From Two Merged Primitive Nervous Systems

A study linked to Stanford Medicine suggests the human brain evolved from the merger of two distinct primitive nervous systems. The research appears to offer new insights into the evolutionary origins of the vertebrate brain. The findings were published or reported in September 2026. The story gained traction on Hacker News, accumulating 29 points and 11 comments. Further details of the methodology and conclusions require access to the full article.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Five Workflow Practices That Improve Results When Using Claude Code

A developer at NRI Dating has shared practical techniques for getting more reliable output from Claude Code, an AI-powered coding assistant. The key improvements came not from crafting clever prompts but from building better processes around the tool, such as providing full feature context rather than isolated instructions. Storing project conventions — including framework choices, naming rules, and off-limits code — in a repository file reduces repeated corrections across sessions. Asking the AI to inspect existing code before making changes helps prevent fast but inaccurate implementations. For complex features, breaking work into staged steps like inspection, proposal, database changes, and UI implementation leads to more controlled and predictable results.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How Google's Protocol Buffers Cuts Payload Size and Latency in Microservices

Protocol Buffers (protobuf) is a language-agnostic binary serialization format originally developed by Google, now widely used in infrastructure tools such as gRPC, Kubernetes, and Cloud Spanner. Unlike JSON or XML, protobuf uses a separate schema definition file and compiles it into strongly typed code, enabling backward-compatible API evolution without breaking existing services. The format delivers roughly 3–10 times smaller payload sizes and significantly lower deserialization latency compared to text-based alternatives. In one fintech deployment processing millions of daily transactions, adopting protobuf reduced bandwidth consumption by 35% and improved P99 latency by 240 milliseconds. Protobuf integrates natively with streaming and batch systems like Kafka, Apache Beam, and Spark, and serves as the foundation of gRPC's HTTP/2-based communication layer.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Stream Claude AI Tokens from AWS Bedrock Using Node.js 22 and SSE

Developers building chat interfaces with large language models like Claude often face sluggish user experiences because apps wait for a complete response before displaying anything. AWS Bedrock's streaming mode addresses this by pushing individual tokens to the browser as soon as they are generated, rather than returning the full response at once. Using Node.js 22's native fetch and a Lambda-backed Server-Sent Events endpoint, developers can wire up this streaming pipeline in minutes with the official AWS SDK package @aws-sdk/client-bedrock-runtime. The streaming flag must be explicitly set to true in the request payload, otherwise Bedrock defaults to returning a single JSON object with the complete response. Developers should note key limitations, including per-minute token rate limits, the need to manually parse raw JSON lines from the stream, and the fact that not all Claude models on Bedrock support streaming.

0 comments Read more at DEV Community
0
ProgrammingGitHub Blog ·

GitHub Podcast Tackles Hot Debates on AI Coding, RAG, and MCP

The latest episode of the GitHub Podcast takes on several trending debates in the AI development community. The hosts explore whether developers should read AI-generated code or simply trust it. The episode also addresses whether Retrieval-Augmented Generation, known as RAG, is becoming obsolete. Additionally, the podcast examines whether GitHub Skills has effectively displaced the Model Context Protocol, or MCP. The episode is published on the official GitHub Blog as part of an ongoing series covering AI and software development topics.

0 comments Read more at GitHub Blog
0
ProgrammingHacker News ·

Qualitative User Research Can Speed Up Product Feedback Loops

A blog post published on nseldeib.com argues that qualitative user research is a valuable tool for building faster feedback loops in product development. The piece suggests that gathering direct user insights can help teams iterate more quickly and make better-informed decisions. The article was shared on Hacker News, where it received 7 points. At the time of reporting, the post had not yet generated any comments on the platform.

0 comments Read more at Hacker News
0
ProgrammingHacker News ·

Rickub Launches as a New Git Platform Claiming Advanced Capabilities

A new version control platform called Rickub has been introduced and shared on Hacker News under the 'Show HN' category. The platform positions itself as an enhanced alternative to existing Git tools, marketing itself as a highly intelligent solution. The submission received 13 points and one comment on Hacker News at the time of reporting. Limited details are publicly available beyond the platform's website at rickub.com. The project appears to be in an early or promotional stage, with community engagement still minimal.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Why the Metric You Track at Work May Not Reflect the Value You Create

A North Star Metric is defined as a single number representing value actually delivered, yet most people default to tracking activity-based numbers that are easier to count and rise more reliably. Using job searching as an illustration, the article contrasts shallow metrics like applications sent or hours spent with more meaningful ones like interviews earned from roles a candidate would genuinely accept. Each step away from raw activity and toward outcomes better captures whether real progress is being made, but outcome metrics can stay flat for weeks even during periods of hard work. The same pattern applies in professional settings, where tickets closed or commits pushed say nothing about whether anything meaningfully improved for users. The piece also warns that any single metric can be gamed unintentionally, making a counterbalancing guardrail metric necessary to keep the primary measure honest.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Razor Syntax Explained: Mixing C# and HTML Before Blazor Enters the Picture

A developer tutorial series has reached its third part, shifting focus from WPF desktop development to web rendering using Razor syntax. Razor allows C# code to be embedded directly into HTML files using the @ symbol, eliminating the need for a separate templating language. The guide demonstrates core Razor features — including expressions, @foreach loops, and @if conditionals — by rendering a read-only version of the same Task Tracker app built in earlier parts. The resulting page displays task data and completion status but intentionally includes no interactive elements. Interactivity such as checkbox state updates without full page reloads is reserved for the next part of the series, which will introduce Blazor, a framework built directly on top of Razor.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Chaos Cypher Turns Obsidian Vaults into AI-Powered Knowledge Graphs

Chaos Cypher is a tool that converts Obsidian Markdown vaults into searchable, AI-powered knowledge graphs without requiring any manual link creation. Users zip their notes and upload the archive, after which the tool extracts entities and relationships from the text using a large language model. The system builds a graph of people, concepts, and projects mentioned across notes — including connections the user never explicitly linked — and enables semantic search and chat with citations back to source notes. Developers can automate the process via a REST API, while command-line users can import vaults with a single terminal command. Existing manual wiki-links in Obsidian are preserved alongside the AI-discovered layer.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Anti-AI Protest Takes Place in Montreal

A protest against artificial intelligence was held in Montreal, drawing participants who voiced opposition to AI technologies. The demonstration was documented and reported on by the Montreal Gazette. The event reflects growing public concern and debate surrounding the rapid expansion of AI systems. Details about the number of attendees and specific demands of the protesters were not fully disclosed in the available report.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Developer builds local tool to inspect and preserve AI coding agent memory and context

A developer has released AgentMemora, a local inspection tool designed to help users understand and manage the context stored by AI coding agents like Claude Code. The tool addresses a common pain point where important decisions, subagent discoveries, and project knowledge become scattered across multiple sessions and difficult to track. AgentMemora organises this information by project, grouping primary sessions, subagent transcripts, durable memory, and instruction files into a single navigable view. It does not replace existing agent memory systems but acts as a read-only inspection and curation layer over locally stored session data. The tool requires no account, API key, or telemetry, and supports exporting a 'Context Capsule' to help developers preserve critical context before starting a new session.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Webhooks, EventBridge, Event Sourcing, CQRS: A Guide to Event-Driven Systems

Event-driven systems communicate by publishing facts about things that have already occurred, allowing producers and consumers to operate independently without tight coupling. The approach broadly splits into two categories: event-driven integration, which covers technologies like webhooks and EventBridge-style APIs for moving events between systems, and event-driven architecture, which includes patterns like Event Sourcing and CQRS for organizing how applications handle events internally. Webhooks are typically the simplest option for notifying external systems, while EventBridge-style platforms are better suited when events need to reach multiple consumers across distributed services. Event Sourcing is valuable when a full history of state changes is a business requirement, whereas CQRS helps when read and write workloads have significantly different performance or scaling needs. These technologies are not mutually exclusive and are frequently combined depending on whether a system requires external integration, internal architectural structure, or both.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Claude Task Master CLI Automates Full Pull Request Lifecycle Using AI Agents

Claude Task Master is an open-source CLI tool built on the Claude Agent SDK that automates the entire pull request workflow without human intervention. The tool follows a plan-work-PR-verify loop, handling everything from task planning and code changes to CI monitoring, review responses, and auto-merging. It persists state so interrupted workflows resume from where they left off, and supports multiple credential profiles to allow parallel task execution across teams. Beyond the CLI, it also exposes a REST API, an MCP server, and webhook integrations for external systems to trigger and monitor tasks. The tool is aimed at teams already using Claude Code who want fully automated, end-to-end PR management for well-defined development goals.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Lane Files Can Stop AI Models From Inventing False README Claims

AI models drafting how-to documentation frequently fabricate unverifiable details such as pip extras, Python version floors, and sample outputs that were never actually executed. A proposed workflow addresses this by generating a machine-readable JSON inventory of public function signatures via AST parsing, which a model can use to draft parameter tables and procedure skeletons. A 'lane file' then restricts the model to only drafting content provable from that inventory, while flagging install instructions, support windows, and example transcripts as human-only responsibilities. The inventory file is hashed in Git so that future README changes can be traced back to a specific symbol set. The core principle is that any undocumented claim remains 'unsigned' until a human author verifies and attaches it to the inventory hash.

0 comments Read more at DEV Community
← NewerPage 1145 of 3221Older →