Vitalii Serbyn — AI Systems Architect https://serbyn.io/blog I design autonomous agent systems, trust-layered orchestration, and production AI infrastructure. en The Arc Since the Six Were Caught https://serbyn.io/blog/the-arc-since-the-six-were-caught https://serbyn.io/blog/the-arc-since-the-six-were-caught Tue, 23 Jun 2026 00:00:00 GMT After fixing six agent failure modes, the system graduated from outcome checking to path checking. A field report on trajectory evals and generalization. The Calibration Ledger: 58 Runs, 93% Pass, n Is Small https://serbyn.io/blog/the-calibration-ledger-58-runs-93-pass-n-is-small https://serbyn.io/blog/the-calibration-ledger-58-runs-93-pass-n-is-small Wed, 17 Jun 2026 00:00:00 GMT 58 runs, ~93% autonomous pass, and why that number is honest evidence — not a reliability proof. A field report on agent evals at solo scale. False Greens: Three Structural Observations https://serbyn.io/blog/false-greens-three-structural-observations https://serbyn.io/blog/false-greens-three-structural-observations Fri, 12 Jun 2026 00:00:00 GMT How fail-closed design, headless execution seams, and converting incidents to fixtures stop AI agents lying about success. A field report from a solo dev-loop. One-Sided Contracts Break Agent Pipelines https://serbyn.io/blog/one-sided-contracts-break-agent-pipelines https://serbyn.io/blog/one-sided-contracts-break-agent-pipelines Mon, 08 Jun 2026 00:00:00 GMT A strict contract enforced on only one side produces false rejections as reliably as a loose one produces false acceptances. A field report from my fleet. QA Is Only as Honest as Its Coverage https://serbyn.io/blog/qa-is-only-as-honest-as-its-coverage https://serbyn.io/blog/qa-is-only-as-honest-as-its-coverage Wed, 03 Jun 2026 00:00:00 GMT A PR merged over a red CI job while dev-loop QA said green. Here's the root cause, the fix, and why partial greens are the most dangerous lies in agent systems. Permissive Parsing Is a False-Green Factory https://serbyn.io/blog/permissive-parsing-is-a-false-green-factory https://serbyn.io/blog/permissive-parsing-is-a-false-green-factory Sat, 30 May 2026 00:00:00 GMT How a silent None from an AI agent registered as a pass — and the fail-closed contract that fixed it. A field report from a production agent fleet. Six False-Greens in a Self-Auditing Agent Pipeline https://serbyn.io/blog/six-false-greens-in-a-self-auditing-agent-pipeline https://serbyn.io/blog/six-false-greens-in-a-self-auditing-agent-pipeline Tue, 26 May 2026 00:00:00 GMT Six ways my autonomous dev-loop reported success when nothing had landed — and the fail-closed gates that caught each one before they corrupted the ledger.