sdd-explore

Investigates codebases and returns structured exploration analysis for spec-driven development changes.

Updated Mar 7, 2026
One-click install
npx skills add https://github.com/albersg/dotfiles --skill sdd-explore-albersg
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sdd-explore
Source: https://github.com/albersg/dotfiles/tree/main/dotfiles-opencode/.opencode/skills/sdd-explore
Command: npx skills add https://github.com/albersg/dotfiles --skill sdd-explore-albersg

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Before committing to a code change, teams need to understand the current architecture, affected files, and trade-offs between approaches. This Skill performs that investigation as a delegated sub-agent within a spec-driven development (SDD) workflow, returning a structured analysis instead of ad-hoc guesses. ## Core Features & Use Cases - Codebase Investigation: Reads entry points, related functionality, tests, and existing patterns to map the current state and affected areas. - Approach Comparison: Compares multiple implementation approaches with pros, cons, and effort estimates, plus a recommendation and risk list. - Artifact Persistence: Saves an exploration artifact to Engram, OpenSpec, both (hybrid), or returns results only, depending on the configured artifact store mode. - Use Case: An orchestrator agent receives a feature request, delegates exploration to this sub-agent, and gets back a structured report covering current state, affected files, approach options, and readiness for a formal proposal. ## Quick Start Ask the orchestrator to explore how a proposed feature would affect the codebase before drafting a change proposal.

Frequently Asked Questions about sdd-explore

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

FAQPage Schema
How do I explore a codebase before making a change in an SDD workflow?▼

Delegate the topic to the sdd-explore sub-agent through the orchestrator. It investigates relevant code, compares approaches, and returns a structured analysis with a recommendation and risks before any proposal is drafted.

What does the sdd-explore skill return to the orchestrator?▼

It returns a fixed Markdown structure covering current state, affected areas with file paths, compared approaches with pros and cons, a recommendation, risks, and a ready-for-proposal verdict.

Can sdd-explore modify code during exploration?▼

No. It is read-only against the codebase and may only create an exploration.md file inside the change folder when a named change is provided. It never modifies existing files.

Which artifact storage modes does sdd-explore support?▼

It supports engram, openspec, hybrid, and none modes. Engram saves under sdd/{change-name}/explore, OpenSpec follows the openspec convention files, hybrid does both, and none returns results only.

Why should the orchestrator not run sdd-explore instructions inline?▼

The skill is executor-only by design. The orchestrator must delegate to the dedicated sdd-explore sub-agent so exploration runs in isolation and returns a concise structured summary rather than consuming orchestrator context.