fix(experiments): require a winning variation when stopping an iteration

[hsadhvani]

Summary

The launchdarkly-experiment-setup skill told the model it could stop an experiment iteration with "no winner declared." In practice LaunchDarkly requires you to declare a winning variation to stop an iteration, so that guidance led to failed stops.

This updates the skill so that stopping always declares a winningTreatmentId (and winningReason). When an experiment is inconclusive, the guidance is to declare the baseline/control as the winner — there is no "stop without a winner" path.

Paired with a fix to the stop-experiment-iteration MCP tool in gram-functions that makes winningTreatmentId required.

Testing approaches

Skill-only doc change; reviewed against the actual stop-experiment-iteration tool behavior.

Feature flags

N/A

[github-actions]

Skill eval results

Skill	Before	After	Δ
agentcontrol/configs-create	100/100 (4/4)	75/100 (3/4)	-25
agentcontrol/configs-update	80/100 (4/5)	80/100 (4/5)	no change
agentcontrol/configs-variations	80/100 (4/5)	80/100 (4/5)	no change
agentcontrol/tools	75/100 (3/4)	75/100 (3/4)	no change
feature-flags/launchdarkly-flag-create	100/100 (3/3)	100/100 (3/3)	no change

Only suites whose source actually changed since their last recorded score were re-run. Soft-failing while we stabilise the baseline.