DeadCode 👻 — The Ghost of Forgotten Code (Reviving Abandoned Repositories with Local AI)

Description and use

DeadCode is a local AI tool that analyzes abandoned Python repositories, identifies technical debt, explains architecture, and provides a phased revival roadmap, all without sending code to external services.

Problem

Abandoned repositories contain forgotten intentions, orphaned TODOs, undocumented architectures, and unpinned dependencies, making revival difficult.

Outcome

DeadCode enables developers to analyze abandoned Python repositories locally, identify technical debt, understand architecture, and receive a structured revival plan without exposing code to external services.

Workflow

• Point DeadCode at a local folder or upload a .zip archive
• DeadCode performs deterministic static analysis on Python ASTs
• Calculates a Repository Health Heuristic (0–100)
• Interrogates the local AI model via Ollama for explanations and contribution suggestions
• Generates "The Last Will of This Repository" as Markdown and JSON with a phased roadmap

Evidence

DeadCode acts as a digital ghost residing locally inside the repository

Source

Performs deterministic static analysis: Parses Python Abstract Syntax Trees (AST), finds empty function stubs, unearths TODO/FIXME/HACK markers, audits unpinned dependencies, and checks for dangerous anti‑patterns (eval, subprocess shell=True, unsafe pickle.loads).

Source

Interrogates the Ghost: Uses a locally running, quantized open‑weight model via Ollama to explain unfamiliar functions, decipher architectures, and suggest contribution starting points—with line‑numbered citations.

Source