Build an Artist-Facing Python Tool
The brief
Write a tool that solves one real pipeline annoyance — asset naming validation, batch export, material auditing, or scene cleanup. It must have a dry-run mode, never destroy data silently, and produce a readable report.
Starter assets
Pick a task you have personally done by hand more than five times. That is your tool. Write the report format before the logic.
Performance budget
- Runs over 100 assets in under 30 seconds.
- Dry-run mode is the default.
- Every destructive action is logged with the affected asset path.
Engine instructions
Milestones
Write the report first
Define exactly what the tool will tell the artist: per-asset status, the rule that failed, and the suggested fix. Build this as plain text output before any UI.
Implement rules as data
Express each validation rule as data (name, description, check function, severity) rather than hard-coded ifs, so rules can be added without touching the engine of the tool.
Dry run and apply
Add an apply mode that performs fixes, with dry run as the default. Log every change with the asset path and the before/after value.
Minimal UI
Add a small UI: rule checkboxes, a run button, a results list, and an apply button that is disabled until a dry run has been done.
Failure handling
Handle locked files, missing dependencies and unexpected asset types without aborting the whole batch. Report failures as results, not as a traceback.
Rollout plan
Write how the tool gets to the team: where it lives, how it is versioned, who owns it, and how someone reports a bug. Then get one artist to use it.