chore: drop redundant Catalog column from README skills table

[mosheabr]

Onboarding type

• New product onboarding (new `components.d/.yml` file)
• Other (catalog change, README fix, infrastructure, etc.)

Other context

The Catalog column in the Available Skills table is redundant and was hiding skills under aggregated rows. Removing it.

Why

• The Skills count column already tells you how many skills are under each product
• The Source column already links to the upstream source repo where skills live
• For single-skill products the Catalog cell was just a duplicate pointer
• For multi-skill synced products (cuOpt 12, NeMo MBridge 20, NemoClaw 10, etc.) the cell pointed at one arbitrary primary catalog_dir while the rest of the skills were invisible at the table level
• After PR chore(regenerate-readme): aggregate rows by product name #200's aggregation work landed for Physical AI, the asymmetry surfaced as a visible bug: row says "Physical AI | 7 skills" but the Catalog cell only listed 6 of 7 (manual side listed all 5, synced side listed only the primary)

Changes

• Drop the Catalog column from both the table header and per-row output in `regenerate-readme.sh`
• Drop `catalog_cell` from the structured TSV emitted by both the synced and manual loops, and from the awk aggregation pass
• Regenerated `README.md` reflects the column removal across all rows (local re-runs without `VERSIONS_FILE` produce em-dash version cells — the sync workflow always populates that file in production, so version cells will repopulate on the next sync)

Verified locally

• Physical AI row: now renders as `| Physical AI | ... | 7 | [Source] | — |` — matches the actual 7-skill consolidated count
• Other rows (cuOpt, NeMo MBridge, NemoClaw, etc.): drop one column each, no other change
• Help table: untouched

All PRs

• All commits signed off with DCO (`git commit -s`).