extract-system-spec

Generate an anonymized, standalone rebuild specification from an entire codebase using multi-agent analysis.

Updated May 3, 2026
One-click install
npx skills add https://github.com/Claudfather/clauDNA --skill extract-system-spec-claudfather
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: extract-system-spec
Source: https://github.com/Claudfather/clauDNA/tree/main/skills/extract-system-spec
Command: npx skills add https://github.com/Claudfather/clauDNA --skill extract-system-spec-claudfather

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When you need to port, hand off, or rebuild a system without the original repo, manually documenting every behavior, data model, and integration is slow and error-prone. This Skill orchestrates a coverage-gated, multi-agent investigation that produces a single standalone Markdown specification an independent team could implement from scratch. ## Core Features & Use Cases - Scoped extraction: Run against the full system, a single functional wing, or a targeted slice (e.g. auth and billing flows), with a Phase 0 cost estimate and budget gate before any heavy work begins. - Six-phase method: Complete inventory, per-wing dossiers, cross-cutting investigations (flows, data lineage, interfaces, security, UI), whole-system reconciliation, spec authoring, and an adversarial review panel covering coverage, reconstruction, flexibility, privacy, and consistency. - Hard anonymization: The deliverable contains no PII, secrets, paths, symbols, or project identity, making it suitable for vendor-neutral handoff, due diligence, escrow, or disaster-recovery documentation. - Use Case: You are re-platforming a legacy proprietary system to a new stack. Run the skill at full-system scope to produce an anonymized rebuild blueprint with acceptance criteria, data semantics, and known unknowns that a fresh team can implement without repo access. ## Quick Start Ask the agent to run the extract-system-spec skill on this repository at full-system scope with a token budget, then confirm the Phase 0 cost estimate to begin the extraction.

Frequently Asked Questions about extract-system-spec

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

FAQPage Schema
How do I create a rebuild specification for an existing codebase?

Run the extract-system-spec skill at the repository root, choosing full-system, wing, or targeted scope. It orchestrates subagents through inventory, dossier, cross-cutting, reconciliation, drafting, and adversarial review phases to produce one standalone Markdown specification.

What is the difference between extract-system-spec and a code audit?

extract-system-spec answers how to rebuild the system elsewhere and outputs an anonymized reconstruction spec. An audit answers what is wrong with the system and outputs ranked findings with fixes for a codebase you intend to keep.

Can I extract a spec for only part of a system?

Yes. Use the wing scope to cover one functional capability plus its seams, or the targeted scope with a quoted description such as auth and billing flows. Scope changes coverage breadth, never the thoroughness of the in-scope analysis.

Does the generated specification contain sensitive or identifying information?

No. The deliverable is hard-anonymized: no PII, secrets, organization identity, source paths, code symbols, or repository metadata. A dedicated privacy review gate and optional exhaustive re-scan verify this before completion.

When should I not use full-system spec extraction?

Avoid it for finding bugs in a system you are keeping, understanding a single module, or documenting small projects. Full-system extraction is deliberately exhaustive and can consume hundreds of thousands of tokens, so confirm the Phase 0 cost estimate first.