johari-diagnostic

To audit an AI agent codebase for observability gaps, run a diagnostic scan that detects tracing, evals, provenance, and red-teaming signals across 16 dimensions. It scores coverage across four quadrants and outputs a formal gap report with actionable remediation recommendations.

The Agentic Johari Window is a four-quadrant model—Open, Blind Spot, Hidden, and Unknown—used to evaluate agent observability. It categorizes signals from your codebase to identify which areas of tracing, evals, and red-teaming are actively monitored versus completely undetected.

You score observability coverage by detecting tracing and red-teaming signals across 16 codebase dimensions and applying weighted multipliers to each of the four Johari quadrants. This produces a scored gap report that highlights your largest observability weaknesses.

Yes, the diagnostic tool identifies the target codebase language and ecosystems before scanning for observability signals. It applies the four-quadrant scoring universally, making it suitable for evaluating any AI agent project with tracing, evals, or provenance data.

The observability diagnostic report includes a concise inline summary and a formal written report. It highlights your largest coverage gaps across the 16 dimensions and provides concrete, actionable remediation steps to improve your agent's tracing and red-teaming signals.

Unknown observability blind spots occur when an AI agent project lacks sufficient tracing, provenance, or red-teaming signals in the codebase. A diagnostic scan maps these missing signals to the Unknown quadrant, quantifying the gap and recommending concrete remediation steps.