Add --show-matrix to print a scenario's appropriateness matrix

[bamdadd]

Context

Contributors authoring or debugging a scenario have no way to see its matrix without reading scenarios.py. A read-only --show-matrix that prints, per attribute × recipient, whether the flow is forbidden / appropriate / unlisted-allowed makes scenarios self-documenting. This is display only — it must not run an agent or touch scoring.

Acceptance criteria

• --show-matrix in src/context_leak/cli.py, honouring --scenario <id> (defaults to the club-reserve scenario) and exiting 0.
• Output shows each (attribute, recipient) cell classified as forbidden, appropriate, or allowed, using recipient roles for readability.
• No agent is run and no score is computed on this path.
• Works alongside --json for a machine-readable matrix.
• A test asserting a known forbidden cell and a known appropriate cell appear correctly for a built-in scenario.