Vibe Code Catalog Commands SOPs Market Radar GitHub

Vibe Code

Benchmarked, versioned workflows for AI coding. No unproven claims.

Get new packs + re-verification alerts

One email when a pack ships or gets re-benchmarked on a new model. No noise.

Claude Code Skills & Subagents

Packs in progress — see the Market Radar for what this category rewards.

Agent Loops & Orchestration

Agent Loops Starter — one loop, one file, green tests or it keeps going

Use this pack when you want a coding agent to keep working until a measurable condition is met instead of stopping after one attempt — failing tests to green, lint clean, build pas

free v0.1.0
Claw Starter — one model, one memory, one schedule, one chat surface

Use this pack to graduate from sessions to a claw — a persistent, scheduled, memory-holding agent that runs standing duties and remembers between runs. Portable across Claude Code,

v0.1.0
Loop Patterns — the four loops of loop engineering, installable

Use this pack when you want to engineer the loop instead of babysitting the agent — the four loop types (agent, verification, event-driven, hill-climb) as installable skills. Keywo

v0.1.0
Window Surfer — /goal: "do 4 hours of work", and mean it

Use this pack when you pay for a Claude plan with rolling usage windows and want none of it wasted — set a timed goal, let the agent work to budget, checkpoint, and auto-resume whe

free v0.1.0

Cursor Rules & Workflows

Packs in progress — see the Market Radar for what this category rewards.

Spec & PRD Templates for AI Coding

Spec Loop Pro — from vague idea to drift-checked feature, 4 skills

Use this pack when features built by AI agents drift from what you asked for — vague ideas become vague code. Spec-first workflow: interview → spec → review gate → task decompositi

v0.1.0

MCP Server Starters

Packs in progress — see the Market Radar for what this category rewards.

Eval Suites & LLM Testing

Packs in progress — see the Market Radar for what this category rewards.

Agent Teams & Role Bundles

Ship Team Starter — 3 native subagents + 1 command, no conversion scripts

Use this pack when one agent doing everything produces sloppy features — you want distinct roles with a review gate: architect designs, implementer builds, reviewer blocks. Keyword

free v0.1.0

Prompt Systems & Goal Loops

Packs in progress — see the Market Radar for what this category rewards.