Vibe. Laugh. Code.https://computergeneratedreality.com/https://computergeneratedreality.com/posts/loop-engineering-epic-loop-skill/https://computergeneratedreality.com/posts/loop-engineering-epic-loop-skill/Loop engineering treats the agent loop, not the prompt, as the thing you design. I distilled the idea into a Claude Code skill called /epic-loop. Here is the principle behind it, why it's worth trying even if you don't want autonomous agents, and the full skill to copy.Fri, 03 Jul 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/behavioral-firewalls-for-ai-agents-compiling-tool-call-telemetry-into-a-finite-a/https://computergeneratedreality.com/posts/behavioral-firewalls-for-ai-agents-compiling-tool-call-telemetry-into-a-finite-a/LLM agents invoke external services through tool-call protocols like MCP. Today's firewalls intercept these calls, validate schemas, and scan signatures. Each call is judged alone. The paper challenges that assumption. An adversary injecting instructTue, 12 May 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/filtering-non-actionable-static-analysis-warnings-with-sentence-transformers-sta/https://computergeneratedreality.com/posts/filtering-non-actionable-static-analysis-warnings-with-sentence-transformers-sta/Between 35% and 91% of warnings produced by static code analysis (SCA) tools are non-actionable — a range documented across multiple prior studies and tools. Non-actionable includes both false positives, where the tool flags something that is not an Tue, 12 May 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/recursive-agent-optimization-training-a-model-to-call-itself/https://computergeneratedreality.com/posts/recursive-agent-optimization-training-a-model-to-call-itself/Recursive Agent Optimization is a reinforcement learning method for training language model agents that delegate sub-tasks to new instantiations of themselves. The question: can a model be taught to call itself, decide what to hand off, and combine tTue, 12 May 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/claude-architect-certification-cheat-sheet/https://computergeneratedreality.com/posts/claude-architect-certification-cheat-sheet/A visual cheat sheet for the Claude Certified Architect Foundations exam, optimized for pattern recognition. Covers agentic orchestration, tool design, prompt engineering, context management, and the decision patterns the exam rewards.Fri, 10 Apr 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/inside-ai-agent-architecture-dispatch-loop/https://computergeneratedreality.com/posts/inside-ai-agent-architecture-dispatch-loop/A technical breakdown of how AI coding agents work internally: the dispatch loop, context window management, tool call execution, error recovery strategies, and what separates demo-quality agents from production-ready ones.Thu, 09 Apr 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/150-claude-agents-same-data-different-results/https://computergeneratedreality.com/posts/150-claude-agents-same-data-different-results/A deep analysis of a new paper that deployed 150 autonomous Claude Code agents on identical NYSE data. The agents diverged on methodology, exhibited model-specific empirical styles, and converged through imitation rather than understanding when shown exemplar papers.Mon, 23 Mar 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/human-certified-module-repositories-ai-supply-chain/https://computergeneratedreality.com/posts/human-certified-module-repositories-ai-supply-chain/A paper proposes Human-Certified Module Repositories — vetted, provenance-rich package ecosystems for an era where AI agents assemble code from open source components.Mon, 23 Mar 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/swe-ci-can-ai-agents-maintain-codebases/https://computergeneratedreality.com/posts/swe-ci-can-ai-agents-maintain-codebases/The first benchmark built on the continuous integration loop. 100 tasks across real repos with 233-day evolution histories and 71 consecutive commits. What it reveals about agent maintenance capabilities.Mon, 23 Mar 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/vibe-code-bench-58-percent-accuracy/https://computergeneratedreality.com/posts/vibe-code-bench-58-percent-accuracy/A deep dive into Vibe Code Bench, the first benchmark that tests whether AI models can build complete web applications from a text spec. 16 frontier models tested, 100 app specs, 10,131 substeps. The best model passes 62% of user workflows. The data reveals exactly where AI-generated apps break.Mon, 23 Mar 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/vibepass-ai-models-cannot-find-their-own-bugs/https://computergeneratedreality.com/posts/vibepass-ai-models-cannot-find-their-own-bugs/Salesforce researchers tested 12 frontier LLMs on self-diagnosing and repairing their own coding faults. The results: 86% of generated test inputs are syntactically valid, but only 61% actually expose bugs. Self-generated tests that miss the fault make repair worse than not testing at all.Mon, 23 Mar 2026 00:00:00 GMThttps://computergeneratedreality.com/posts/the-wrong-benchmark-why-human-level-misses-what-actually-matters-in-ai-refactori/https://computergeneratedreality.com/posts/the-wrong-benchmark-why-human-level-misses-what-actually-matters-in-ai-refactori/Mon, 16 Mar 2026 08:59:19 GMThttps://computergeneratedreality.com/posts/needs-title-research-a-practical-guide-for-establishing-a-technical-debt-managem/https://computergeneratedreality.com/posts/needs-title-research-a-practical-guide-for-establishing-a-technical-debt-managem/Fri, 13 Mar 2026 08:25:23 GMThttps://computergeneratedreality.com/posts/handling-tokenization-structured-inputs/https://computergeneratedreality.com/posts/handling-tokenization-structured-inputs/Tue, 25 Nov 2025 00:00:00 GMThttps://computergeneratedreality.com/posts/deploying-vesuvian/https://computergeneratedreality.com/posts/deploying-vesuvian/Thu, 13 Nov 2025 09:52:47 GMThttps://computergeneratedreality.com/posts/cgr-labs-day-4/https://computergeneratedreality.com/posts/cgr-labs-day-4/Fri, 24 Oct 2025 00:00:00 GMThttps://computergeneratedreality.com/posts/exploring-droid/https://computergeneratedreality.com/posts/exploring-droid/Mon, 20 Oct 2025 19:15:13 GMThttps://computergeneratedreality.com/posts/cgr-labs-day-3/https://computergeneratedreality.com/posts/cgr-labs-day-3/Wed, 08 Oct 2025 14:39:33 GMThttps://computergeneratedreality.com/posts/cgr-labs-day-2/https://computergeneratedreality.com/posts/cgr-labs-day-2/Mon, 06 Oct 2025 00:00:00 GMThttps://computergeneratedreality.com/posts/cgr-labs-day-1/https://computergeneratedreality.com/posts/cgr-labs-day-1/Sun, 05 Oct 2025 00:00:00 GMThttps://computergeneratedreality.com/posts/hetzner-k3s/https://computergeneratedreality.com/posts/hetzner-k3s/Sun, 05 Oct 2025 00:00:00 GMT