Separate CHASM scheduler action capacity APIs

[chaptersix]

What changed?

• Split the CHASM scheduler action-capacity predicate and mutation into canTakeScheduledAction and consumeScheduledAction.
• Update only the direct query and consumption call sites.
• Add focused tests for query non-mutation and zero, one, and multiple successful consumptions.

Production behavior is unchanged. Each successful limited-action consumption decrements RemainingActions exactly once and increments the conflict token exactly once; N successful consumptions preserve an N-token delta. Paused or exhausted attempts mutate neither counter, and unlimited consumption remains allowed without changing either counter.

Why?

This is PR 1 of the CHASM scheduler readability stack. It removes the boolean mutation mode from useScheduledAction(bool) so callers state whether they are querying capacity or consuming it, without introducing the lifecycle classifier, transition helpers, buffer planner, or later refactors.

Stack

• PR 0: Characterize CHASM scheduler task validity #35, main <- sch-readable
• PR 1 (this PR): Separate CHASM scheduler action capacity APIs #36, sch-readable <- sch-readable-action-capacity

How did you test it?

• built
• run locally and tested manually
• covered by existing tests
• added new unit test(s)
• added new functional test(s)

.bin/gci-v0.13.6 write --skip-generated -s standard -s default <changed Go files>
go test -tags test_dep ./chasm/lib/scheduler -run <focused action-capacity and affected behavior tests> -count=1
go test -tags test_dep ./chasm/lib/scheduler -count=1
GOLANGCI_LINT_BASE_REV=sch-readable GOLANGCI_LINT_FIX=false make lint-code
git diff --check

Potential risks

The risk is limited to accidentally changing action-limit or conflict-token accounting while separating the APIs. The focused tests pin the paused, unlimited, exhausted, single-consumption, and multi-consumption behavior and exact counter/token deltas.