acquire-codebase-knowledge

Generates seven evidence-backed documentation files mapping a codebase's stack, structure, and architecture.

Updated Jun 14, 2026
One-click install
npx skills add https://github.com/ironkid90-s/lucky5-v7 --skill acquire-codebase-knowledge-ironkid90-s
Or copy as Structured Prompt for Agentā–¼
Please help me install this Agent Skill.
Skill: acquire-codebase-knowledge
Source: https://github.com/ironkid90-s/lucky5-v7/tree/main/.github/skills/acquire-codebase-knowledge
Command: npx skills add https://github.com/ironkid90-s/lucky5-v7 --skill acquire-codebase-knowledge-ironkid90-s

SYSTEM DOCUMENTATION & REQUIREMENTS

šŸ’” This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve? Onboarding into an unfamiliar repository requires hours of manual exploration, and README files often describe intended rather than actual architecture. This Skill automates repository discovery by scanning manifests, configs, git history, and source files to produce verifiable documentation grounded only in what the code actually shows. ## Core Features & Use Cases - Automated Repository Scanning: Runs a Python script that detects manifests for 25+ languages, CI/CD pipelines, containers, security configs, code metrics, TODOs, and high-churn files. - Seven Structured Documents: Populates STACK.md, STRUCTURE.md, ARCHITECTURE.md, CONVENTIONS.md, INTEGRATIONS.md, TESTING.md, and CONCERNS.md in docs/codebase/ using templates with required evidence citations. - Validation Loop: Enforces an output contract where every claim is traceable to source, unknowns are marked [TODO], and intent-dependent decisions are flagged as [ASK USER] questions. - Use Case: A developer joining a legacy .NET and Godot project asks to map the codebase; the Skill scans the repo, documents the actual layered architecture, and surfaces tech debt and fragile high-churn files. ## Quick Start Ask the AI to map this codebase and generate onboarding documentation in docs/codebase/.

Frequently Asked Questions about acquire-codebase-knowledge

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

FAQPage Schema
How do I document an existing codebase automatically?ā–¼

Run the scan.py script from the project root, then follow the four-phase workflow to populate seven templates in docs/codebase/. Every claim must cite evidence from source files, configs, or terminal output rather than assumptions.

What languages and frameworks does codebase scanning support?ā–¼

The scanner detects manifests for over 25 languages including Node.js, Python, Go, Rust, Java, .NET, Ruby, PHP, and Elixir. It also identifies CI/CD pipelines, Docker configs, and monorepo tools like pnpm workspaces, Nx, and Turborepo.

What are the requirements to run the codebase scan script?ā–¼

The scan requires Python 3.8 or later and git installed on the system. Run it from the target project root using only Python standard library modules, so no additional package installation is needed.

Can I document only part of a codebase, like architecture or testing?ā–¼

Yes, focus area mode lets you fully complete specific documents first, such as architecture only or testing and concerns. Phase 1 scanning still runs in full, and non-focus documents keep required sections with unknowns marked as [TODO].

How does the skill handle outdated or misleading README files?ā–¼

The workflow cross-references README claims against actual file structure before treating them as fact. Intent-vs-reality divergences are explicitly highlighted in the final summary alongside numbered [ASK USER] questions.

When should I not use automated codebase documentation?ā–¼

Avoid it for routine feature implementation, bug fixes, or narrow code edits where repository-level discovery is unnecessary. It is designed for explicit mapping, documentation, and onboarding requests only.