VibeFlow
CLAUDE.md8 starsChanged 8 months agoMITView on GitHub
VibeFlow is a workflow template that enforces docs-first, TDD-driven development for AI-assisted software engineering. It uses Claude Code skills and hooks to ensure documentation stays current and tests drive implementation. All documentation MUST exist and be validated BEFORE any code is written. Stages A-E produce documents; Stages F-H implement code that must match those documents. These rules are deterministic — enforced by hooks and skills: Workflow stages, tracks, branch conventions, file naming, and skills are documented in .claude/rules/. See README.md…
In this repository10 other files it gives its agents.
- analyze-codebaseSkill
- clarify-demandSkill
- create-feature-specSkill
- define-prdSkill
- define-tech-specSkill
- manage-workSkill
- prepare-releaseSkill
- record-decisionSkill
- run-tddSkill
- validate-checkpointSkill
The file
We split this CLAUDE.md into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 15
Branch binding (UserPromptSubmit hook): All prompts blocked unless current branch is feat/<slug> for an active work item.
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
VibeFlow — AI-Assisted Docs-First Development Workflow
Project Identity
VibeFlow is a workflow template that enforces docs-first, TDD-driven development for AI-assisted software engineering. It uses Claude Code skills and hooks to ensure documentation stays current and tests drive implementation.
Core Principle: Docs-First Mandate
All documentation MUST exist and be validated BEFORE any code is written. Stages A-E produce documents; Stages F-H implement code that must match those documents.
Enforcement Rules
These rules are deterministic — enforced by hooks and skills:
- Branch binding (UserPromptSubmit hook): All prompts blocked unless current branch is
feat/<slug>for an active work item. Manage-work commands are exempted so users can register new work items from main. - Checkpoint gates (two layers):
- UserPromptSubmit hook:
checkpoint-gate.pyblocks prompts containing "advance"/"close" if checkpoint not passed (deterministic safety net) - Manage-work skill:
advanceandclosecommands validate checkpoints before updating the manifest (skill-instructed, redundant backup)
- UserPromptSubmit hook:
- Exception: Prompts are always allowed when no manifest exists (initial project setup) or when no active work items are present
- All hooks are read-only — no hook mutates any file. All manifest updates happen in skills.
Reference
Workflow stages, tracks, branch conventions, file naming, and skills are documented in .claude/rules/. See README.md for the full pipeline diagram and contributor details.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

