Suggest edit — frameworks

Title

Name

Note

---
visibility: public
---

# frameworks & methodologies

multi-skill systems that define how you work with agents. these aren't individual tools — they're development philosophies packaged as skills.

## obra/superpowers

[github.com/obra/superpowers](https://github.com/obra/superpowers) — **147k+ stars.** the leading agentic skills framework. defines the methodology many others follow.

**skills included:**
- `test-driven-development` — enforces TDD, 85-95% coverage
- `systematic-debugging` — structured debugging
- `verification-before-completion` — prevents premature task completion
- `brainstorming` — structured idea → design transformation (see [[communication-and-writing]])
- `writing-plans` / `executing-plans` — plan creation and execution
- `dispatching-parallel-agents` — multi-agent parallelization
- `requesting-code-review` / `receiving-code-review` — two-stage review
- `using-git-worktrees` — isolated worktrees
- `finishing-a-development-branch` — branch completion workflow
- `subagent-driven-development` — subagent dispatch with review checkpoints
- `root-cause-tracing` — trace errors to their origin (see [[data-and-analysis]])
- `writing-skills` / `using-superpowers` — meta-skills

**why it's the best:** consistently high quality, battle-tested by thousands, enforces good practices (TDD, code review, verification) rather than just adding capabilities. many individual skills appear in [[development]] and [[security]].

also: **[superpowers-marketplace](https://github.com/obra/superpowers-marketplace)** (821 stars) — curated plugin marketplace.

## everything-claude-code

[github.com/affaan-m/everything-claude-code](https://github.com/affaan-m/everything-claude-code) — **151k stars.** agent harness optimization system: skills, instincts, memory, security, research-first development. recommended in the [vibe coding manifesto](https://www.moonflowers.xyz/writing/vibe_coding_manifesto).

## NeoLab context-engineering-kit

[github.com/NeoLabHQ/context-engineering-kit](https://github.com/NeoLabHQ/context-engineering-kit) (796 stars) — hand-crafted skills for improving agent output quality.

- `software-architecture` — Clean Architecture, SOLID, design patterns
- `subagent-driven-development` — dispatch with review checkpoints
- `prompt-engineering` — Anthropic best practices, agent persuasion
- `kaizen` — continuous improvement (Japanese/Lean methodology)

## SuperClaude Framework

[github.com/SuperClaude-Org/SuperClaude_Framework](https://github.com/SuperClaude-Org/SuperClaude_Framework) (22.2k stars) — config framework with specialized commands, cognitive personas, and dev methodologies.

## get-shit-done (TACHES)

[github.com/gsd-build/get-shit-done](https://github.com/gsd-build/get-shit-done) (50.7k stars) — meta-prompting, context engineering, spec-driven development.

## HumanLayer commands

[github.com/humanlayer/humanlayer](https://github.com/humanlayer/humanlayer/tree/main/.claude/commands) — the research/plan/implement command set recommended in the [vibe coding manifesto](https://www.moonflowers.xyz/writing/vibe_coding_manifesto):
- `research_codebase`
- `create_plan`
- `implement_plan`
- `create_handoff`
- `resume_from_handoff`