chore: populate computed data for secondary leaves in multi-charm products

[srbouffard]

Context

The migration in #7 converted existing computed/*.json files to the new leaf_metrics envelope by mapping the old flat metrics to the first inline leaf of each root product. This was correct for single-charm products (Discourse → discourse-k8s, NetBox → netbox-k8s, etc.).

For multi-charm products, only the first leaf has real metrics — the remaining leaves currently show insufficient_data:

Root product	First leaf (has data)	Secondary leaves (no data yet)
Matrix	synapse	saml-integrator
Wazuh	wazuh-server	wazuh-indexer, wazuh-dashboard
Indico	indico-operator	indico-nginx-ingress (if applicable)

What needs to happen

This resolves itself automatically once the nightly scorer runs against these products. When make score PRODUCT=matrix or the nightly GHA workflow runs, all leaf units will be scored and their metrics written to the leaf_metrics envelope.

No code changes are needed — this is a tracking issue to confirm secondary leaf data is populated after the next nightly run following merge of #7.

Acceptance criteria

• After the first nightly run post-merge, computed/matrix.json has metrics for both synapse and saml-integrator
• After the first nightly run post-merge, computed/wazuh.json has metrics for wazuh-server, wazuh-indexer, and wazuh-dashboard
• The UI shows real medals (not insufficient_data) for all composed leaves of the above products

Notes

• No action needed before merge — this is a post-merge data verification task
• If the nightly run fails to populate secondary leaves, the scorer contract in logic.py or the GHA invocation should be checked