Add Databricks notebook for installing Genie Code skills

[Praneeth16]

Summary

• Adds a Databricks notebook (databricks-skills/install_genie_code_skills.py) that installs Genie Code skills from within a workspace, no local terminal needed
• Uses the Databricks Python SDK (pre-installed in all runtimes) to download skills from GitHub and upload them to /Users/<username>/.assistant/skills/
• Works on any compute including serverless
• Skill lists and extra file mappings match install_skills.sh exactly (validated programmatically)
• Updates README with a "From a Databricks notebook" section under Genie Code Skills

Test plan

• Python syntax validation passes (ast.parse)
• All imports resolve (urllib, json, posixpath, databricks.sdk)
• Skill lists match install_skills.sh: 26 Databricks + 8 MLflow + 1 APX
• Extra files mappings match install_skills.sh for all 20 Databricks + 5 MLflow + 1 APX entries
• GitHub raw URLs tested for Databricks, MLflow, and APX skills
• Ran notebook on fe-vm-cme-conde workspace: 35 skills installed successfully in 59 seconds
• Verified /Users/<username>/.assistant/skills/ contains all 35 skill directories via databricks workspace list
• No emdashes, AI slop, or hyperboles in any text

This pull request was AI-assisted by Isaac.

[scottDBX1886]

I like the approach for V1. I think we need to make an update for v2 to dynamically generate the install skills list instead of hardcoded in the notebook as this introduces 2 places (NB and install script) where we have to maintain the skill install lists.

One change first though.

The verify cell is not working. Please see the below change I had to make. The current cell gives "Could not list skills: 'ObjectInfo' object has no attribute 'is_directory'" when executed.

from databricks.sdk import WorkspaceClient

w = WorkspaceClient()
username = w.current_user.me().user_name
skills_path = f"/Users/{username}/.assistant/skills"

try:
entries = list(w.workspace.list(skills_path))
subdirs = sorted([e.path.split("/")[-1] for e in entries if str(e.object_type) == "ObjectType.DIRECTORY"])
if subdirs:
print(f"Found {len(subdirs)} sub-directories in {skills_path}:\n")
for name in subdirs:
print(f" {name}")
else:
print(f"No sub-directories found in {skills_path}.")
except Exception as e:
print(f"Could not list sub-directories: {e}")

[Praneeth16]

@scottDBX1886 thanks for the suggestions! I've implemented them and made the skills discovery dynamic based on GitHub.

[scottDBX1886]

New notebook looks really good. Dynamic skills retrieval is perfect. Cell 7 looks good.
Very nicely done.

[scottDBX1886]

All looks good. great PR