bmad-walkthrough

Guides human reviewers through code changes via orientation, walkthrough, risk pass, and testing steps.

Updated Sep 4, 2026
One-click install
npx skills add https://github.com/Joao-Victor-Medeiros/save-sc --skill bmad-walkthrough-joao-victor-medeiros
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: bmad-walkthrough
Source: https://github.com/Joao-Victor-Medeiros/save-sc/tree/main/.agents/skills/bmad-walkthrough
Command: npx skills add https://github.com/Joao-Victor-Medeiros/save-sc --skill bmad-walkthrough-joao-victor-medeiros

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Reviewing a code change without guidance is slow and error-prone: reviewers miss the intent behind a diff, overlook high-risk spots, and lack a structured way to verify behavior. This Skill walks a human reviewer through a change step by step, from understanding its purpose to making an approve/rework decision. ## Core Features & Use Cases - Orientation and Intent Summary: Identifies the change from a PR, commit, branch, or spec, then presents the intent and surface-area stats (files changed, modules touched, boundary crossings). - Concern-Based Walkthrough: Organizes the diff by design concerns with clickable path:line stops, generating a review trail automatically when the author did not provide one. - Risk and Testing Passes: Surfaces high-blast-radius spots (auth, schema, billing, security) and suggests concrete manual observations to verify behavior. - Use Case: A teammate opens a PR touching authentication middleware. Trigger the walkthrough to get the intent summary, a guided tour of the key code locations, a list of risk spots like token validation changes, and suggested manual tests before approving. ## Quick Start Ask the assistant to walk me through this change or give me a human review of the current PR.

Frequently Asked Questions about bmad-walkthrough

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

FAQPage Schema
How do I review a pull request with an AI walkthrough?▼

Trigger the walkthrough skill and pass a PR reference, which it resolves via gh pr view. It then presents the intent, a concern-organized tour of path:line stops, risk spots, and testing suggestions before asking for your approve or rework decision.

How does the walkthrough identify what change to review?▼

It checks in order: an explicit argument (PR, commit, branch, spec), recent conversation context, sprint status files for stories in review, and the current git HEAD. If none identify a change, it asks you directly and halts after three failed exchanges.

What happens if the change has no spec or review trail?▼

The skill generates a review trail from the diff by reading changed files in full, identifying 2-5 design concerns, and selecting path:line stops for each. If git is unavailable and trail generation fails, it falls back to a non-trail walkthrough mode.

Can I customize the walkthrough workflow steps?▼

Yes, through TOML overrides in _bmad/custom/ that merge with the skill's customize.toml. You can add activation steps before or after greeting, inject persistent facts, and define an on_complete action run after the review decision.

Does the walkthrough replace automated tests and CI?▼

No, it explicitly assumes CI and test suites exist and work. Its testing step focuses on manual observation of user-visible behavior, such as UI changes, CLI output, and API responses, that automated checks do not cover.