Command Claude Code
Deep-analyzes the codebase to discover patterns, conventions, and architecture. Builds .vibeflow/ knowledge base.
Spec-driven development for AI agents. Define what to build, let the agent build it right. Claude Code · Cursor · Copilot.
This repository also configures its own agents. See what vibeflow tells them →
Command Claude Code
Deep-analyzes the codebase to discover patterns, conventions, and architecture. Builds .vibeflow/ knowledge base.
Command Claude Code
Audits implementation against its DoD and project patterns, from a review package of the diff. Runs tests, tests scenarios, sabotages risk code, scans the diff for destructive ops (Critical Gate), and reports PASS / PARTIAL / FAIL with a severity and a cause on every finding. With --reaudit, checks a fix round against…
Command Claude Code
Runs an interactive discovery dialogue (1-5 rounds) to turn a vague idea into a clear, actionable PRD.
Command Claude Code
Generates a technical spec with DoD, scope, anti-scope, and applicable patterns from .vibeflow/.
Command Claude Code
Fixes an observed defect with reproducible evidence in one call: trace doc before code, the fix, and a regression test written before the fix.
Command Claude Code
Implements a feature from its spec with guardrails: budget, DoD, anti-scope, and pattern compliance.
Command Claude Code
Generates a self-contained prompt pack from a spec with embedded patterns for any coding agent.
Command Claude Code
Fast-tracks small tasks (≤4 files) into a prompt pack in one command. Skips discover, generates ephemeral spec.
Command Claude Code
Compiles audit statistics: pass/fail rates, most violated patterns, and quality trends.
Command Claude Code
Updates .vibeflow/ with corrections, new conventions, decisions, or patterns. Also imports from external repos via --from.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: