Software engineering often assumes that developers can operate at full capacity indefinitely. Fatigue, illness, stress, and poor recovery can reduce concentration, code review quality, and decision-making, increasing the risk of defects and rework. Most developer tools focus on the code and leave the developer to manage this human context separately. Po-Ko, inspired by the ancient Greek word *ptochos*, represents the strength of acknowledging human limits and the need for support. It combines wearable signals such as overnight HRV, sleep, and stress with short self-reported context to produce an explainable 1-10 capacity score. The score shows why support may be needed instead of presenting an unexplained health label. Designed for software engineers, Po-Ko makes IBM Bob the primary interaction surface. Through a local Model Context Protocol server, a developer can check their current score, log daily metrics, submit context that sensors cannot capture, and receive time-specific advice. The React/Vite dashboard provides a visual view with score breakdowns, metric explanations, and recovery trends. Po-Ko is local-first: telemetry is stored in SQLite on the developer's machine. Its learning engine can analyse personal telemetry and logged sick-day history for recurring pre-sick patterns, while the Code Guardian connects a high-risk support state to Git commit metadata and review guidance. This helps the developer make a safer work decision before fatigue-related mistakes become expensive rework, without blocking the developer or exposing biometric data to a cloud service. What makes Po-Ko unique is that it connects personal capacity to the software workflow instead of treating health and development as separate systems. It is efficient because IBM Bob retrieves the context, interprets the signals, and carries the support state into code review, reducing manual copying, missed warnings, and avoidable rework.
Category tags: