bama
BAMA: Build Any Mac App. Agent skills to research, clone, design, build, optimize, and test native macOS apps with Swift and SwiftUI.
翻譯準備中
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證A Claude Code plugin that adds an agentic skills framework, with MCP's and a meta prompting processes. Inspired by "Superpowers" & "Figma Console MCP", but not just for development or design; but for all aspect within the digital production cycle.
翻譯準備中
取自儲存庫 README
正在載入完整 README。
檢視完整 README ↗One pack of agent skills for digital production work: 283 skills across 16 domains, from design and research to legal, security and accounting. Each skill is a plain SKILL.md folder that any agent supporting the Agent Skills format can load.
What you get:
design-empathize, development-plan, ...). Each says what it needs first, what it does step by step, and where it writes its result.workflows/design.yml, workflows/launch-product.yml, ...) that puts skills into ordered phases. The workflow skill runs one phase at a time and stops for your approval at every boundary.scripts/gate.sh, plus a write hook in Claude Code). Pass --skip-checks to go ahead anyway; the skip is recorded..singularskills/log.jsonl, so you, your team and the singular-status skill can see where each project stands.Migrating from metapowers: move
.metapowers/to.singularskills/and replace/domain:skillwith/domain-skill. Artifacts written before the run log existed count as written: the conductor offers them for approval instead of re-running their skills.
Any agent (the universal installer copies skills/ into each detected agent's skills folder):
npx skills add bromso/singularskills # this project
npx skills add bromso/singularskills --global # every project, user scope
npx skills add bromso/singularskills --agent codex -y # one agent, no prompts
Skills land in .agents/skills/ for Codex, OpenCode, Cursor, GitHub Copilot, Gemini CLI and Kimi Code CLI, and in .claude/skills/ for Claude Code. Each skill keeps its scripts/ (the phase gates, executable) and references/ folders.
Claude Code (also installs the /design-empathize-style command wrappers and the gate hook):
/plugin marketplace add bromso/singularskills
/plugin install singularskills@singularskills
Codex:
codex plugin marketplace add bromso/singularskills
codex plugin add singularskills@singularskills
GitHub Copilot CLI and Gemini CLI: use the universal installer. For Gemini command wrappers (/design-empathize and the rest), also copy adapters/gemini/commands/*.toml into .gemini/commands/ (project) or ~/.gemini/commands/ (user).
Workflow artifacts are written to .singularskills/<domain>/<project>/ in your project. That folder is meant to be committed: it is the record of what each phase produced, and the gates read it.
Then try a skill (the Claude Code and Gemini CLI command form; other agents pick skills up by name or description):
/design-empathize date-picker
/research-discover user-onboarding
/development-plan auth-system
Tier 1, tested per release (scripts/smoke-install.sh installs into each and runs a gate): Claude Code, Codex, OpenCode, Cursor, GitHub Copilot (VS Code and CLI), Gemini CLI, Kimi Code CLI.
Tier 2, supported through the universal installer, not tested per release: everything npx skills add supports, including Windsurf, Zed, Cline, Antigravity, Roo Code, Qwen Code, Goose, Amp, Kilo Code, Trae and the rest.
| Domain | Methodology | Skills | Description |
|---|---|---|---|
| Design | Design Thinking | 9 | Component design from research to usability test |
| Research | Double Diamond | 4 | Structured discovery and validation |
| Development | Dev Workflow | 6 | Plan, build, test, debug, review, ship |
| Marketing | RACE Framework | 37 | Full growth marketing lifecycle |
| Branding | Brand Lifecycle | 6 | Brand creation and compliance auditing |
| Accessibility | WCAG-EM | 6 | Accessibility audit and remediation |
| Project Management | Scrum | 32 | Sprint-based project management |
| Coaching | Domain Coaching | 5 | Expert feedback across 5 disciplines |
| Leadership | Leadership Dev | 12 | Team building and management tools |
| Legal |
Each domain's skills follow the phases of a proven methodology. A workflow file lists the phases, the artifact each one writes, and what it requires:
name: design
folder: design
phases:
- phase: 1
title: Empathize
artifact: 01-empathize.md
requires: []
skills: [design-empathize]
- phase: 2
title: Define
artifact: 02-define.md
requires: [01-empathize.md]
skills: [design-define-component]
/design-empathize date-picker writes .singularskills/design/date-picker/01-empathize.md./workflow design date-picker runs phase 1, plays back what it decided, and asks before it moves to phase 2./design-define-component date-picker will not write 02-define.md until 01-empathize.md exists, unless you pass --skip-checks..singularskills/<folder>/<project>/. Commit them./workflow launch-product acme # Run a cross-domain workflow phase by phase
/singular-status # Where every project stands, from the run log
/singular-report # Cross-domain summary report
/singular-detect # Scan the project's stack, suggest skills
/singular-find-skill <topic> # Search skill marketplaces
/singular-create-skill <topic> # Write a new skill in this pack's format
/singular-improve-skill <name> # Refine a skill from feedback
Full documentation: bromso.github.io/singularskills
singularskills/
├── skills/ # The pack: one folder per skill (SKILL.md, scripts/gate.sh, references/)
├── workflows/ # One YAML file per workflow: phases, artifacts, requirements
├── references/ # Reference docs shared by several skills; stacks/ for project stacks
├── evals/ # claude plugin eval cases, one folder per skill (bun run evals)
├── adapters/ # Per-agent wrappers: Claude commands, gates.json and hook; Gemini commands
├── .claude-plugin/ # Claude Code plugin manifest and marketplace
├── packages/
│ ├── pack-tools/ # Pack checks and generators (gates, references, adapters)
│ └── skill-validator/ # SKILL.md validation
├── apps/docs/ # Documentation site (Fumadocs + Next.js)
├── scripts/ # run-evals.sh, smoke-install.sh, check-rename.sh
└── docs/ # Design specs and implementation plans
git clone https://github.com/bromso/singularskills.git
cd singularskills
bun install
bun run build # Build all packages
bun run test # Run all tests
bun run lint:skills # Validate every SKILL.md
bun run check:pack # Pack checks: workflows, generated gates/references/adapters, counts
bun run evals --dry # List the eval cases (free); bun run evals runs them (spends credits)
bun run check # Biome check
bun run format # Biome format
See AGENTS.md for how the repository fits together.
See CONTRIBUTING.md for development guidelines.
MIT — Jonas Broms
| Legal Lifecycle |
| 35 |
| Contracts, compliance, governance |
| Security | NIST CSF 2.0 | 39 | Cybersecurity framework |
| Compliance | GRC Lifecycle | 51 | 30+ regulatory frameworks |
| Bookkeeping | Bookkeeping Cycle | 9 | Transaction recording and reporting |
| Accounting | Accounting Cycle | 13 | Full 8-phase accounting cycle |
| Finance | FP&A | 11 | Financial planning and analysis |
| Singular | Cross-domain | 8 | The workflow conductor, plus status, init, report, detect, find, create and improve skills |