Fix CI runner startup_failure after org transfer

[rayketcham]

CI runs show startup_failure since the repo was transferred from rayketcham to rayketcham-lab. The self-hosted runner needs to be registered to the org or the repo needs a runner attached.

[rayketcham]

Root Cause

The self-hosted runner Ubuntu2 is registered to rayketcham-lab org in the Default runner group, but that group likely doesn't include project-forge in its allowed repositories.

Fix (requires org admin in GitHub UI)

1. Go to https://github.com/organizations/rayketcham-lab/settings/actions/runner-groups
2. Click on the Default runner group
3. Under "Repository access", add project-forge
4. Save

Alternatively, run:

gh auth refresh -h github.com -s admin:org
# Then:
gh api orgs/rayketcham-lab/actions/runner-groups/1/repositories -f selected_repository_ids[]=$(gh api repos/rayketcham-lab/project-forge --jq '.id')

Verified

• Runner Ubuntu2 is active and serving other org repos (iron-temple works)
• CI workflow is correctly configured with runs-on: [self-hosted, Linux]
• 154 tests pass locally, 0 warnings

[rayketcham]

Resolved — CI runners fixed across rocky, Ubuntu2, ubuntu3 with dynamic Python 3.12+ detection (FIND_PY), venv setup, and ruff formatting alignment.