Red — one phase, manually driven

This is a manual phase control, not a separate method. It runs step 1 and step 2 of the predictive cycle and then stops. Use it when you want to steer Red-Green-Refactor yourself instead of invoking the whole exact-coding workflow. The method lives in one place; this skill only selects which part of it to execute now.

What to read

1. .pi/skills/predictive-tdd/SKILL.md — sections Core rules, 1. Activate one behavior and 2. Reach behavioral Red. Those are binding as written.
2. The matching profile under .pi/skills/predictive-tdd/stacks/ for inactive-test syntax and the commands to run.
3. .pi/skills/exact-coding-shared/human-in-the-loop.md for the Red checkpoint and the prediction-mismatch rule.

If no ordered test list exists yet, say so and offer test-list first. Do not invent a list as a side effect of this phase.

What to do

Execute exactly those two steps of the cycle, then stop:

• activate one behavior from the test list,
• predict the narrowest test's outcome, failure category, and message or values,
• run it and compare prediction with reality,
• add only enough scaffold to reach a behavioral failure, re-predicting each time.

Where to stop

Stop at the Red checkpoint from .pi/skills/exact-coding-shared/human-in-the-loop.md and report the active behavior, the prediction, the actual result, and why this failure is the intended behavioral Red. Do not write production behavior, and do not continue into Green — even when the Autonomy Level would not stop there. The human invoked a single phase; that invocation is the checkpoint.

If the activated test already passes, confirm that prediction, record that no production change is needed, and stop. Do not manufacture a failure.

On a prediction mismatch, follow Prediction mismatch in the Predictive TDD skill and stop there.