Application developers don't always scan each package before incorporating it, and now coding agents install packages on their own, which makes install time the first control point well before a scanner ever sees the code. Cycode is providing early access to Workstation Protection, which applies two controls in real time, checking packages against a threat intelligence feed and blocking anything updated too recently to have been properly vetted. See why package policy is now an agent governance decision: buff.ly/1gFQkn9 #DevSecOps #SupplyChainSecurity #AppSec #Cycode

Talentica Software this week launched DevX AI Pods, a managed software delivery service that uses AI agents trained to supply the context needed to deploy applications built with AI coding tools at scale. The agents evaluate code against a CCCR framework while more than 600 Talentica engineers validate the output, addressing a gap where AI tools generate more code without actually increasing the rate at which applications ship. Read the full article on the service and the survey data behind the code volume problem: buff.ly/exiItJq #AI #DevOps #SoftwareDevelopment

AWS has added a harness to its open source Strands SDK that gives developers a fully assembled, customizable AI agent as a foundation, complete with shell, file and web tools, long-term memory and a context window capability that can cut token consumption by as much as 28%. In effect, it is now possible to create a working agent with one line of code, and as it becomes easier to build agents, the focus of DevOps teams turns more toward deployment, though that pace will still be limited by what they can observe, control and prove. Read the full article on the Strands harness and what agentic engineering at scale will demand: buff.ly/JeGXW0M #AWS #AIAgents #OpenSource #DevOps

🏆 Nominations are officially open for the 10th Annual DevOps Dozen Awards. For a decade, the DevOps Dozen has recognized the people, projects, tools and platforms making an impact on software delivery. This year features 24 award categories across two divisions: 🔹 9 Community Awards — free to enter 🔹 15 Tools & Services Awards — $300 per nomination Know a leader, team, project or solution that deserves recognition? Send in your nomination. 📅 Nominations close October 18, 2026 at 5 PM ET. 👉 Submit your nomination: devopsdozen.com #DevOps #DevOpsDozen2026 #DevOpsAwards #SoftwareDelivery

Splunk plans to release an open source AI model trained to analyze log data, helping DevOps teams reason across growing volumes of telemetry more efficiently. Read more: buff.ly/s4zgSJT #Splunk #Observability #OpenSource #DevOps #OpenTelemetry

Splunk has open sourced Token Meter, a tool that reads local Claude Code and Codex logs to surface live cost and activity for an AI agent session, pricing trace files against public model rates and flagging long-running tasks that could run on a cheaper model. It arrives as a Futurum Group survey finds just under half of organizations are over their AI budgets, with token volumes climbing even as the cost of an individual token keeps falling. Read the full article on Token Meter and the push to keep AI coding costs under control: buff.ly/v7eEDU6 #Splunk #Cisco #AI #DevOps #AICosts

Existing DevOps workflows designed for humans will not scale when developers are soon creating 15 to 20 pull requests a week, which is the premise behind Harness previewing a revamped Software Factory platform built around four core AI agents handling delivery, security testing, runtime security and cost management on top of a knowledge graph now tracking 567 entity types and 29 billion context updates a month. Harness also previewed Vibe Mode for citizen developers and a Flex Pricing option with shared credits across modules, though it remains unclear how quickly teams will actually push agents into their pipelines. Read the full article on the agentic engineering era and what it means for DevOps teams: buff.ly/Tn2OCVw #DevOps #AIAgents #Harness #PlatformEngineering