v2.0.0: floop-client — Review Loop Only

🎉 v2.0.0 — Major Refactor

Highlights

fdesign extracted to standalone project. All prototype, design token, component, sitemap, and journey-map logic has been moved to the new fdesign project. floop-client is now focused exclusively on its core mission: the review loop.

What Changed

• Removed fdesign prototype commands (token, component, sitemap, journey check, prototype init/validate)
• Removed src/floop/prototype.py, src/floop/tokens.py — now in fdesign
• Removed docs/ mkdocs site — replaced by standalone README
• Added src/floop/project.py — local project/version container management
• Added tests/test_project.py — 100% coverage on project module
• Refactored src/floop/cli.py — slimmed from 1752 to ~900 lines
• Refactored src/floop/skills.py — review-only skill prompts
• Updated README with problem/solution narrative, use cases, quick start
• Added LICENSE (MIT)

Responsibility Boundary

floop-client (v2.0.0)	fdesign (new)
.floop/ workspace	.fdesign/ workspace
Project & version containers	Design tokens (W3C DTCG)
floop review upload	Component library (YAML DSL)
floop comments	Sitemap & journey map
floop resolve	fdesign journey check
floop preview	fdesign preview

Tests

32 passed, 100% coverage.

floop 1.1.0 - Feedback Skill Loop

floop 1.1.0 - Feedback Skill Loop

This release introduces a complete feedback collection and iteration loop for floop prototypes.

Highlights

• Added the new floop-feedback Agent Skill for collecting and analyzing reviewer comments from floop-server.
• Added floop feedback CLI support for fetching comments from published review versions.
• Split prototype publishing and feedback collection into clearer workflows: floop-prototype builds and publishes, floop-feedback monitors reviewer comments.
• Improved Agent Skill guidance so agents know how to fix feedback, create a new floop version, upload it, and share the updated review URL.
• Strengthened review setup guidance around floop review set so agents configure review publishing before upload.

Feedback Workflow

The intended loop is now explicit:

1. Build or update a prototype.
2. Publish it with floop review.
3. Share the generated review URL with reviewers.
4. Fetch reviewer comments with floop feedback.
5. Apply feedback, create a new version, and publish again.

Quality

• Test suite remains at 100% coverage.
• 543 tests passing at release time.

1.0.0

AI prototypes degrade with every iteration. floop is the missing quality loop that keeps your Agent honest. An open-source alternative to Figma Make and Google Stitch.*