https://newruntime.com/feeds/coding-agents.xmlNew Runtime: Coding agentsA living map of coding agents as they move from autocomplete into supervised software work, review loops, environments, and operating cost. This bounded feed exists because the topic is a featured, evidence-dense New Runtime hub.2026-09-01T00:00:00.000ZNew Runtimehttps://newruntime.com/New Runtime tracks the systems, signals, and shifts reshaping software as agents become first-class participants in work, products, and organizations.https://newruntime.com/patterns/verification-bandwidth-is-the-scarce-resource/Verification bandwidth is the scarce engineering resource2026-09-01T00:00:00.000Z2026-09-01T00:00:00.000ZThe primary bottleneck in agentic software delivery is moving from code production to the human and machine capacity required to verify it.https://newruntime.com/projects/coding-agent-cost-controls/Coding Agent Cost Controls2026-08-15T00:00:00.000Z2026-08-15T00:00:00.000ZA forkable setup for reducing coding-agent cost by shaping environment output, repo maps, model routing, cache stability, and task-level budgets.https://newruntime.com/projects/verifiable-ai-coding-workflow/Verifiable AI Coding Workflow2026-08-06T00:00:00.000Z2026-08-06T00:00:00.000ZA small agent-assisted coding workflow where the useful artifact is not generated code, but a checked path from idea to tests, diff, review, and handoff.https://newruntime.com/posts/agentic-sdlc-software-factory-loop/A Software Factory Connects Agents Through Verified Outcomes2026-08-03T00:00:00.000Z2026-08-03T00:00:00.000ZAugment and Warp describe team-level agent loops that move work from trigger and specification through implementation, verification, release, and measured improvement.https://newruntime.com/posts/claude-code-auto-mode-action-gate/Claude Code Auto Mode Gates Actions Instead Of Explanations2026-08-03T00:00:00.000Z2026-08-03T00:00:00.000ZClaude Code Auto Mode combines an input injection probe with a two-stage action classifier, preserving autonomy while exposing an honest residual miss rate.https://newruntime.com/posts/cline-hooks-agent-harness-guardrails/Cline Hooks Put Deterministic Rules Inside The Agent Loop2026-08-03T00:00:00.000Z2026-08-03T00:00:00.000ZCline's plugin hooks show how an agent harness can journal every run and block dangerous tool calls without waiting for the model to choose a guardrail.https://newruntime.com/posts/evocode-bench-multi-turn-regressions/EvoCode-Bench Exposes Multi-Turn Regression Risk2026-08-03T00:00:00.000Z2026-08-03T00:00:00.000ZEvoCode-Bench tests coding agents across persistent workspaces and evolving requirements, where regressions become the dominant failure mode.https://newruntime.com/posts/google-adk-loop-engineering-state-pruning/Loop Engineering Needs State Pruning, Not Infinite Chat2026-08-03T00:00:00.000Z2026-08-03T00:00:00.000ZThe Google ADK loop-engineering article frames self-correcting agents as desired-state systems with pruning, validation, and circuit breakers.https://newruntime.com/shifts/coding-agent-general-workbench/Coding tool -> general-purpose workbench2026-08-01T00:00:00.000Z2026-08-01T00:00:00.000ZCoding agents are expanding beyond software implementation into context-aware workbenches that assemble prototypes, interfaces, documents, workflows, and operational artifacts.https://newruntime.com/posts/openai-gpt-5-6-price-performance-frontier/GPT-5.6 Turns Efficiency Work Into API Economics2026-08-01T00:00:00.000Z2026-08-01T00:00:00.000ZOpenAI turned GPT-5.6 serving and kernel efficiency gains into lower Luna and Terra prices, plus a faster Sol mode for latency-sensitive API workloads.https://newruntime.com/posts/langchain-reviewbench-review-agent-evals/ReviewBench Turns Code Review Into An Agent Eval2026-08-01T00:00:00.000Z2026-08-01T00:00:00.000ZLangChain's ReviewBench uses real PR review history to test whether code-review agents can recover substantive reviewer findings without flooding humans with noise.https://newruntime.com/posts/github-stacked-prs-agent-review/GitHub Stacked PRs Turn Large Agent Changes Into Reviewable Chains2026-07-31T00:00:00.000Z2026-07-31T00:00:00.000ZGitHub's stacked pull request preview gives large dependent code changes a native review path, which matters as agents produce broader diffs.https://newruntime.com/posts/cline-recursive-self-improvement-coding-agent/Cline Turns Recursive Self-Improvement Into Harness Work2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZCline's Terminal-Bench run is not a singularity story; it is a concrete loop where an agent reads traces, patches the harness, reruns evals, and hands a PR to humans.https://newruntime.com/posts/cursor-cloud-agent-environment-product/Cursor Treats the Cloud Agent Environment as the Product2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZCursor's cloud-agent environment write-up shows why agent performance depends on dependencies, commands, security boundaries, end-to-end tests, and self-healing diagnostics.https://newruntime.com/posts/cursor-agent-swarms-sqlite/Cursor's SQLite Swarm Makes Coordination the Expensive Part2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZCursor's SQLite experiment shows why agent-swarm economics depend on task trees, shared memory, conflict handling, review lenses, and selective use of expensive planners.https://newruntime.com/posts/factory-comarch-night-shift-agents/Factory and Comarch Show the Night-Shift Shape of Agent Work2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZFactory's Comarch case study is less about one coding assistant and more about governed agent missions that continue execution while humans set direction and review.https://newruntime.com/signals/factory-at-comarch-hundreds-of-engineers-now-use-the-factory-platform-to-run-age/Factory: At Comarch, hundreds of engineers now use the Factory platform to run agents end-to-end across...2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZA public X post from Factory with a linked primary source flags At Comarch, hundreds of engineers now use the Factory platform to run agents end-to-end across the software development lifecycle. • 40% greater engineering efficiency • >30% high...https://newruntime.com/signals/github-stacked-prs-now-on-github/GitHub: Stacked PRs now on GitHub 🥞2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZA public X post from GitHub with a linked primary source flags Stacked PRs now on GitHub 🥞https://newruntime.com/signals/google-gemini-create-edit-and-summarize-with-gemini-on-macos-using-your-voice/Google Gemini: Create, edit, and summarize with Gemini on macOS using your voice.2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZA public X post from Google Gemini with a linked primary source flags Create, edit, and summarize with Gemini on macOS using your voice. Stay in your flow by speaking into any active window: Dictate clean text or ask Gemini to transform highlighted...https://newruntime.com/signals/mem0-experiment-2-does-a-preference-survive-a-hard-context-wipe/mem0: Experiment 2: Does a preference survive a hard context wipe?2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZA public X post from mem0 with a linked primary source flags Experiment 2: Does a preference survive a hard context wipe? We told Claude(mid-conversation) to write functions with standard for loops instead of list comprehensions. Then ran /...https://newruntime.com/posts/minisqlite-makes-the-swarm-claim-inspectable/miniSQLite Makes the Coding-Swarm Claim Inspectable2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZThe released Rust database turns Cursor's swarm experiment into an artifact that can be read, built, tested, and challenged instead of accepted as a benchmark chart.https://newruntime.com/signals/openai-developers-imagegen-in-codex-just-got-a-new-lightbox-and-canvas/OpenAI Developers: ImageGen in Codex just got a new lightbox and canvas.2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZA public X post from OpenAI Developers with a linked primary source flags ImageGen in Codex just got a new lightbox and canvas. Now, it’s even easier to explore and refine visuals in your workflow.https://newruntime.com/signals/openai-we-re-also-upgrading-auto-review-in-the-chatgpt-app-and-codex-cli-from-gp/OpenAI: We’re also upgrading Auto-review in the ChatGPT app and Codex CLI from GPT-5.4 to GPT-5.6 Luna.2026-07-30T00:00:00.000Z2026-07-30T00:00:00.000ZA public X post from OpenAI with a linked primary source flags We’re also upgrading Auto-review in the ChatGPT app and Codex CLI from GPT-5.4 to GPT-5.6 Luna. Combined with Luna’s new price, we expect Auto-review to cost about 10x less, makin...https://newruntime.com/posts/codex-multi-agent-orchestration-skill/A Codex Skill Turns Multi-Agent Work Into a Reusable Control Surface2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZA shared Codex orchestration skill points to a future where agent workflows are packaged as portable operational routines.https://newruntime.com/signals/codex-multi-agent-orchestration-skill/A Codex Skill Turns Multi-Agent Work Into a Reusable Control Surface2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZA shared Codex orchestration skill points to a future where agent workflows are packaged as portable operational routines.https://newruntime.com/posts/agentic-engineering-full-stack-workflow/Agentic Engineering Looks Like Workflow Design, Not Hands-Free Coding2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZA full-stack agentic engineering walkthrough shows the operating pattern around planning, validation, browser checks, and human review.https://newruntime.com/signals/agentic-engineering-full-stack-workflow/Agentic Engineering Looks Like Workflow Design, Not Hands-Free Coding2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZA full-stack agentic engineering walkthrough shows the operating pattern around planning, validation, browser checks, and human review.https://newruntime.com/signals/augment-code-the-best-value-for-every-token-gpt-5-6-sol-is-now-our-default-in-co/Augment Code: The best value for every token: GPT-5.6 Sol is now our default in Cosmos Eight models in eight...2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZA public X post from Augment Code as a public source in its own right flags The best value for every token: GPT-5.6 Sol is now our default in Cosmos Eight models in eight weeks! After testing them all on real, long-horizon software development tasks, we’r...https://newruntime.com/posts/claude-code-certification-credential-layer/Claude Code Certifications Turn Agent Use Into a Credential Layer2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZAnthropic certification activity around Claude Code is a signal that agentic development is becoming a managed enterprise capability.https://newruntime.com/signals/claude-code-certification-credential-layer/Claude Code Certifications Turn Agent Use Into a Credential Layer2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZAnthropic certification activity around Claude Code is a signal that agentic development is becoming a managed enterprise capability.https://newruntime.com/posts/claude-md-prompt-cruft-shrinkage/Claude.md Shrinkage Says the Harness Is Learning What Not to Say2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZA reported 80 percent reduction in Claude Code prompt material points to a maturing harness discipline around concise system context.https://newruntime.com/signals/claude-md-prompt-cruft-shrinkage/Claude.md Shrinkage Says the Harness Is Learning What Not to Say2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZA reported 80 percent reduction in Claude Code prompt material points to a maturing harness discipline around concise system context.https://newruntime.com/shifts/verification-ownership/Code production -> verification ownership2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZAs agents produce more implementation, engineering responsibility is moving toward specifications, evidence, acceptance decisions, and release accountability.https://newruntime.com/posts/code-review-becomes-agent-bottleneck/Code Review Becomes the Agent Bottleneck2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZColes shows that reviewing generated code is becoming the scarce engineering work, not merely a cleanup pass after agent output.https://newruntime.com/signals/code-review-becomes-agent-bottleneck/Code Review Becomes the Agent Bottleneck2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZColes shows that reviewing generated code is becoming the scarce engineering work, not merely a cleanup pass after agent output.https://newruntime.com/posts/codex-hooks-close-the-type-error-loop/Codex Hooks Close the Type Error Loop2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZCodex hooks make validation output an active part of the agent loop, so type errors and checks can be returned before the task drifts.https://newruntime.com/signals/codex-hooks-close-the-type-error-loop/Codex Hooks Close the Type Error Loop2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZCodex hooks make validation output an active part of the agent loop, so type errors and checks can be returned before the task drifts.https://newruntime.com/posts/codex-security-cli-scan-workbench/Codex Security CLI Turns Security Review Into a Scannable Workbench2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZOpenAI's Codex Security CLI packages repository, diff, working-tree, export, validation, and patch flows into a security-review workbench rather than a single scanner command.https://newruntime.com/signals/cursor-cursor-is-now-on-ipad/Cursor: Cursor is now on iPad.2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZA public X post from Cursor with a linked primary source flags Cursor is now on iPad. All the power of Cursor on iPhone, with more room to work with agents.https://newruntime.com/signals/openai-developers-we-used-gpt-5-6-sol-in-codex-to-optimize-its-own-infrastructur/OpenAI Developers: We used GPT-5.6 Sol in Codex to optimize its own infrastructure and performance.2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZA public X post from OpenAI Developers as a public source in its own right flags We used GPT-5.6 Sol in Codex to optimize its own infrastructure and performance. These improvements compound across inference and the agent loop, producing more useful work from...https://newruntime.com/signals/openai-we-quietly-released-the-open-source-codex-security-cli-but-hacker-news-fo/OpenAI: We quietly released the open-source Codex Security CLI, but Hacker News found it before we had...2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZA public X post from OpenAI with a linked primary source flags We quietly released the open-source Codex Security CLI, but Hacker News found it before we had a chance to share it here... You can now use it to scan repositories, track findings...https://newruntime.com/posts/prompt-caching-turns-context-into-infrastructure/Prompt Caching Turns Agent Context Into Infrastructure2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZEarendil frames prompt caching as an agent systems primitive, where stable context becomes a cost, latency, and architecture concern.https://newruntime.com/signals/prompt-caching-turns-context-into-infrastructure/Prompt Caching Turns Agent Context Into Infrastructure2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZEarendil frames prompt caching as an agent systems primitive, where stable context becomes a cost, latency, and architecture concern.https://newruntime.com/posts/raptor-loop-hunt-security-agent-loop/RAPTOR Loop Hunt Packages Security Hunting as an Agent Skill2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZRAPTOR Loop Hunt shows security research moving toward looped agent skills with altitude changes, evidence collection, and review checkpoints.https://newruntime.com/signals/raptor-loop-hunt-security-agent-loop/RAPTOR Loop Hunt Packages Security Hunting as an Agent Skill2026-07-29T00:00:00.000Z2026-07-29T00:00:00.000ZRAPTOR Loop Hunt shows security research moving toward looped agent skills with altitude changes, evidence collection, and review checkpoints.https://newruntime.com/signals/augment-code-our-users-have-been-really-excited-to-try-kimi-k3-and-we-re-excited/Augment Code: Our users have been really excited to try Kimi K3 and we're excited to bring it to Augment's p...2026-07-28T00:00:00.000Z2026-07-28T00:00:00.000ZA public X post from Augment Code as a public source in its own right flags Our users have been really excited to try Kimi K3 and we're excited to bring it to Augment's product family. It is the most capable open-source model that we have tested to date....https://newruntime.com/signals/cursor-today-we-re-launching-cursor-start-a-new-649-month-plan-for-developers-in/Cursor: Today we're launching Cursor Start, a new ₹649/month plan for developers in India.2026-07-28T00:00:00.000Z2026-07-28T00:00:00.000ZA public X post from Cursor with a linked primary source flags Today we're launching Cursor Start, a new ₹649/month plan for developers in India. Start includes generous access to Grok 4.5 and Composer, so you can plan, build, test, and ship...https://newruntime.com/signals/factory-we-re-firm-believers-in-the-open-security-ecosystem-contributing-across/Factory: We're firm believers in the open security ecosystem, contributing across our platform: • Publi...2026-07-28T00:00:00.000Z2026-07-28T00:00:00.000ZA public X post from Factory with a linked primary source flags We're firm believers in the open security ecosystem, contributing across our platform: • Publishing two open-weight models behind Droid Shield 2.0 • Using Autonomous Security Revi...https://newruntime.com/signals/github-5/GitHub: 5.2026-07-28T00:00:00.000Z2026-07-28T00:00:00.000ZA public X post from GitHub with a linked primary source flags 5. Turn on code scanning Code scanning uses CodeQL to identify patterns that lead to vulnerabilities, including injection flaws, unsafe deserialization, and insecure GitHub Action...https://newruntime.com/signals/github-you-re-not-behind/GitHub: You're not behind.2026-07-28T00:00:00.000Z2026-07-28T00:00:00.000ZA public X post from GitHub as a public source in its own right flags You're not behind. There's no secret everyone else has. There's just the harness, and it's mostly all you need. @burkeholland gives you a simple, repeatable workflow for GitHub Co...