Articles, releases and code from Hacker News, Reddit, GitHub and the people building RPA, workflow automation and AI agents — plus what the community pushed to the top today.
AIA hacker breached over 100 enterprise websites using autonomous AI agents, highlighting the need for developers to harden automated systems and APIs against multi-agent cyberattacks.
AIOpenFlow is an open-source visual workflow builder built on the opencode engine that enables developers to orchestrate multi-agent AI systems visually.
AIGoogle Cloud's Agent Registry enables dynamic MCP server discovery, allowing AI agents to query endpoints via URNs rather than relying on hardcoded URLs.
AIBotline is an open-source Cloudflare Worker that provides cross-vendor messaging, file exchange, and webhook doorbells to coordinate AI agents across Anthropic, xAI, and Meta.
AICelorga released an open-source, Emacs-free Org parser and MCP runtime, enabling AI agents to read, manipulate, and execute tasks across plain-text Org files.
AIParallel AI detailed an architecture that exposes 600-plus API endpoints to AI agents via MCP using dynamic OpenAPI generation and a single shared route catalog.
AIn8n published a guide detailing how to run local AI automation workflows on an NVIDIA DGX Spark, eliminating cloud API dependencies for sensitive business data.
AIDistributed systems are shifting toward agentic architectures, requiring automation builders to manage non-deterministic runtime decisions, silent reasoning failures, and dynamic tool selection instead of static service workflows.
AIRunning small language models on-device enables automation developers to build local-first applications that minimize cloud API costs, lower latency, and protect sensitive user data.
AIOpenCode offers an open-source, multi-model harness alternative to Claude Code, letting automation builders avoid subscription caps and use diverse LLMs at the cost of manual configuration.
AICamunda 8.10 Self-Managed adds an in-process restore REST API, allowing teams to automate cluster recovery without broker restarts or manual node configuration.
AIResearchers introduced Thought Blocks, a method enabling AI agents to exchange non-autoregressive latent memory, cutting multi-agent communication latency by over eightfold compared to text exchanges.
AIResearchers introduced Hyperdimensional Symbolic Messages to preserve semantic accuracy across noisy multi-agent communication channels, enabling developers to build more reliable and readable collaborative AI agent systems.
AIMistral released Timestral, lightweight time-series foundation models that run on CPUs, letting developers integrate zero-shot forecasting and anomaly detection directly into automated enterprise workflows and agents.
AIGoogle released EmbeddingGemma 2, an open-weight multimodal embedding model that enables local AI agents to retrieve text, code, images, audio, and video from a unified vector space.
AISimon Willison released llm-tools-datasette 0.1.1, fixing a dependency issue to restore Datasette database querying tools for LLM-driven automation workflows.
AIResearch shows LLM pricing agents can sustain above-cost prices depending on prompt history and interface design, requiring automation builders to carefully configure agent inputs to avoid unintended market collusion.
AIA new multi-agent reinforcement learning framework uses decentralized gossip networks to share reputation data, helping autonomous agents foster cooperation without requiring global information.
AIGsearch is an open-source MCP server that uses an automated Firefox browser to scrape structured Google Search results for AI agents without requiring an API key.
AIThis tutorial shows how to implement session persistence for AI agents, enabling builders to save, resume, and safely recover multi-step conversational workflows across interruptions.
AIResearch reveals that 70% of marketplace AI agent skills carry elevated capabilities while frequently underdeclaring permissions, requiring builders to profile actual code bundles instead of trusting manifests.
AIMonday.com rebuilt its API edge using Go and WunderGraph Cosmo, drastically lowering latency and resource consumption to support high-frequency automated calls from AI agents.
AIAPIStack launched an MCP server providing AI agents with deterministic tools for time zones, currency conversions, business days, and identity validation.
AIA developer built a custom GitHub Actions merge queue to coordinate high-volume parallel pull requests generated by multiple autonomous coding agents on personal repositories.
AIA large-scale evaluation showed specialized sales AI agents matched human staff conversion rates at two percent of the cost, outperforming general-purpose models through better qualification workflows.
AINina Technologies launched a service designed to help AI agents and automation workflows access the web reliably without getting blocked by bot-detection systems.