Clean a project folder with AI
Decide which drafts can go and which originals, attachments and versions need to stay. Review the evidence before choosing recovery storage or permanent removal.
To clean a project folder with AI, begin with one directory and a definition of what must survive. An export called final can still depend on a sidecar file. A second copy in another folder may be the only record of a delivery. The artifact-steward skill looks for evidence of those relationships before treating either file as disposable.
You receive reasons for keeping a file, removing it or leaving the decision open. Identical content can establish duplication, but you still need a surviving copy in the right place. A preview lets you choose between moving a file into recovery storage and authorising permanent deletion.
What it does
- Check which large files account for the directory’s size.
- Compare apparently identical drafts using their full contents.
- Look for project references to a proposed removal.
- Preserve attachments associated with a delivered version.
- Review skipped or failed operations after an interruption.
Install and use
- Choose a local directory and a file-capable agent with Python. Document inspection may use additional software.
- Ask “Which drafts of the exhibition booklet could we remove while keeping the delivered files?”
- Inspect the plan, authorise a method and check its results before treating space as reclaimed.
FAQ
Source
Made by Ring0321. Original skill: artifact-steward, published on GitHub under the MIT licence.
View on GitHub