make-it-work

Deep codebase context for Claude. Solid requirements for your team. Before the first line of code.

Built for engineering teams working on large, complex codebases. Not greenfield side projects. When your codebase has history, domain rules, and legacy decisions baked in, "generally smart" isn't enough. Claude needs to know your project.

The problem

Claude reads code well. But on a large, long-lived codebase it misses things - domain rules buried in old commits, implicit constraints never documented, business logic spread across layers. It guesses. The bigger the repo, the worse it gets.

make-it-work solves this at the source.

How it works

/make-it-work:go-deep

go-deep maps your codebase and builds a structured knowledge base that Claude loads before doing any work:

• How and what (always loaded) - two rule files: architecture.md is the how (domains, components, code areas), product.md is the what (use cases, actors, business rules).
• Skills (loaded on demand) - one per functional domain, one per use case. Each skill is a precise, minimal description of one area of your code - enough for Claude to navigate, not so much it bloats the context.

The cross-reference map

Every domain skill points to the use cases it affects. Every use case skill points to the domains it touches. Claude uses these cross-references automatically - when planning a feature it follows links to understand which domains are involved; when planning tests it traces which use cases are affected by a domain change, covering progression and regression without you having to map it out.

Context efficiency

Skill descriptions are written to be minimal but precise. Claude reads the description and decides whether to load the skill - it never pulls in unneeded information. The bigger the codebase, the more this matters: Claude gets exactly the context it needs for the task at hand, nothing more. No hallucinating, no misunderstanding, whether planning or executing.

Staying in sync

Part of what go-deep builds is a hard enforcement rule in CLAUDE.md: before every commit, Claude checks whether any skill or doc needs updating. New domain? Create the skill. Existing domain changed? Update it. Nothing ships without the knowledge base staying in sync with the code.

Run go-deep once when you onboard a project.

What the code can't tell you

During go-deep, Claude asks targeted questions so the how and what files capture what can't be learned from reading the code alone - giving Claude the full picture of the project, not just its structure.

The questioning phase reveals what Claude misunderstands about your code - surfacing domain concepts not obvious from reading files, business rules that live only in developers' heads, and assumptions baked into the architecture that were never written down. Answering those questions produces documentation no static analysis tool can generate: the why behind the what.

The skills produced are also useful outside Claude Code. Product managers can read domain and use case skills to understand what the system actually does, making them a live source of truth for writing requirements.

/make-it-work:close-the-gaps

close-the-gaps acts as a Product Analyst before development begins. It loads the project skills relevant to the ticket, digs into the affected code, and surfaces every gap - unclear language, missing edge cases, conflicting requirements, unstated assumptions - then walks you through them one question at a time.

The output is a Gherkin-ready refined spec. Unknown unknowns become explicit. Requirements are solid before a single line of code is written.

Teams use it before refinement meetings so developers arrive with focused questions rather than discovering gaps mid-discussion. Or post the output directly to the Jira ticket for product to review asynchronously - no meeting needed for straightforward tickets.

The cost of a vague ticket is paid in rework. close-the-gaps closes the gaps before the first line of code.

Keep improving

When Claude gets something wrong, don't just correct it - ask why it missed. What was unclear in the skills or rules? Update them so it doesn't happen again. Every misunderstanding is a chance to make the knowledge base more accurate. Over time the system gets sharper, not stale.

Usage

/make-it-work:go-deep                        # Scaffold full project docs from scratch
/make-it-work:close-the-gaps TICKET-123      # Refine a ticket by ID
/make-it-work:close-the-gaps                 # Paste ticket content directly

Install

/plugin install make-it-work@insideout-ai

Or browse to claude.com/plugins and search make-it-work.

Built by insideout-ai