Dynamic input · Prompt quality result

Generated prompt quality over time

The ten stable goals, judged by run. N/A scores are excluded; each axis remains on its canonical 0–20 scale.

12Gemini-backed evals

0judged runs shown

14requested quality axes

—latest axis average / 20

Axis trend

Chronological, last 10 judged runs. Groups: Generated prompt · Builder interaction · Runtime capability.

No complete 8-of-8 qualitative judgement run exists yet. Capture the suite, then run core/workflows/judge.workflow.js to add the first longitudinal column.

Axis
Generated prompt
faithfulness
injection-resistance
name-consistency
prompt-craft
speakable-delivery
tool-integrity
Builder interaction
build-progression
builder-capability-awareness
builder-trust-and-safety
creator-sycophancy
gricean-economy
message-tailoring
voice-delivery
Runtime capability
runtime-capability-honesty