sdd-init

Initialize SDD project context, testing capabilities, and skill registry with Engram or OpenSpec persistence.

Updated Aug 25, 2026
One-click install
npx skills add https://github.com/CarlosWilliamsR/SketchOS --skill sdd-init-carloswilliamsr
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sdd-init
Source: https://github.com/CarlosWilliamsR/SketchOS/tree/main/.config/opencode/skills/sdd-init
Command: npx skills add https://github.com/CarlosWilliamsR/SketchOS --skill sdd-init-carloswilliamsr

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Starting spec-driven development in a project requires manually detecting the tech stack, test tooling, and conventions, then wiring up persistence and registries by hand. This Skill automates that bootstrap so SDD context, testing capabilities, and the skill registry are initialized consistently. ## Core Features & Use Cases - Stack and Tooling Detection: Inspects package.json, go.mod, pyproject.toml, CI, and lint/test configs to identify the real stack, test runner, coverage, linter, type checker, and formatter without guessing. - Flexible Persistence Modes: Supports engram, openspec, hybrid, and none modes, writing Engram observations, openspec/config.yaml files, or both. - Skill Registry Generation: Builds .atl/skill-registry.md by scanning user-level and project-level skill directories, deduplicating by name. - Use Case: When starting SDD in a new repository, run the init phase to detect that the project uses vitest with coverage, resolve strict TDD mode, persist testing capabilities, and produce a registry plus a structured initialization envelope pointing to the next /sdd-explore or /sdd-new step. ## Quick Start Ask the agent to run sdd init for this project so it detects the stack, configures persistence, and generates the skill registry.

Frequently Asked Questions about sdd-init

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I initialize SDD in an existing project?

Run the sdd-init phase, which inspects files like package.json, go.mod, and pyproject.toml to detect the stack and test tooling. It then persists context and testing capabilities in your chosen mode and builds the skill registry.

What is the difference between engram, openspec, and hybrid persistence modes?

Engram mode saves context and capabilities only to Engram observations without creating openspec files. Openspec mode writes openspec/config.yaml and file artifacts, while hybrid mode does both. None mode returns detected context without writing SDD artifacts.

How is strict TDD mode determined during initialization?

Strict TDD is resolved from an explicit agent marker or openspec/config.yaml first. If none exists but a test runner is detected, it defaults to true; with no test runner it is set to false with an explanation.

What happens if openspec already exists in my project?

The skill reports what already exists and asks before updating it, rather than overwriting existing openspec content. This prevents accidental loss of prior specifications or configuration.

Which skill directories are scanned when building the skill registry?

The scan covers user-level paths such as ~/.claude/skills and ~/.config/opencode/skills plus project-level paths like .claude/skills and .atl/skills. It skips sdd-*, _shared, and skill-registry entries, preferring project-level skills on name conflicts.