ai-sdlc-sat

Bootstrap your codebase with agentic development tools in a single command.

What It Installs

| Asset | Destination | Purpose | |-------|-------------|---------| | GitHub Copilot instruction files | .github/instructions/ | Always-on coding standards that Copilot applies automatically | | Spec-driven workflow prompts | .github/prompts/ | Agent prompts for analyze → plan → spec → execute workflow | | Agent OS directory structure | .agent-os/ | Structured storage for product docs, specs, and task files |

Usage

# Interactive (recommended)
npx ai-sdlc-sat

# Install to a specific directory
npx ai-sdlc-sat ./my-project

# Install everything non-interactively
npx ai-sdlc-sat --skip-prompts

# Instructions only
npx ai-sdlc-sat --instructions-only

# Prompts only
npx ai-sdlc-sat --prompts-only

# Overwrite existing files
npx ai-sdlc-sat --overwrite

Installed Files

GitHub Copilot Instructions (.github/instructions/)

Instruction files auto-apply to matching file types when GitHub Copilot is active.

| File | Applies To | Contents | |------|-----------|----------| | base.instructions.md | All files | Code quality, security (OWASP), error handling, git | | typescript.instructions.md | *.ts, *.tsx | Type safety, strict mode, common TS patterns | | react.instructions.md | Components, pages, hooks | Hooks rules, performance, composition | | node.instructions.md | Server/API files | Async patterns, security headers, env config | | testing.instructions.md | Test files | AAA pattern, coverage, mocking | | git.instructions.md | All files | Conventional commits, branch naming, PR standards |

Workflow Prompts (.github/prompts/)

Prompts are available in GitHub Copilot Chat. Type / to invoke them.

| Prompt | Purpose | |--------|---------| | analyze-product | Scans codebase and creates .agent-os/product/overview.md | | plan-product | Creates/updates .agent-os/product/roadmap.md with milestones | | create-spec | Creates a feature spec at .agent-os/specs/[feature].md | | execute-tasks | Implements a task from a spec with full acceptance criteria |

Agent OS Structure (.agent-os/)

.agent-os/
├── product/
│   ├── overview.md     ← fill in after running analyze-product
│   ├── tech-stack.md   ← define your approved tech stack here
│   └── roadmap.md      ← created by plan-product
├── specs/              ← feature specs created by create-spec
└── tasks/              ← task breakdown files

Recommended Workflow

1. Analyze — Run analyze-product prompt to document your codebase
2. Plan — Run plan-product to define milestones and features
3. Spec — Run create-spec for each feature you want to build
4. Execute — Run execute-tasks referencing a spec to implement it

Prompt Attribution

Prompts included in this toolkit are adapted from the MIT-licensed AgentOS project by Builder Methods. They have been modified to work with GitHub Copilot's .prompt.md format and optimized for personal development environments.

License

MIT