How does the skill detect a multi-file system?

It looks for signals such as importing from 2+ siblings in the same directory, sharing a database table or JSON resource, or being part of a common lifecycle chain.

What is the 'Read-Before-Edit' rule?

This rule mandates that Claude must read every file listed in a 'Related files' header before making any edits, ensuring it understands the full system context and doesn't break dependencies.

Can I use this with any programming language?

Yes, while optimized for the Nuxt 4 stack via cellm, the file-context logic and header formatting are universal and can be applied to any multi-file codebase.

Does this skill add headers to every file in my project?

No. It specifically avoids 'noise' by only adding headers to files that form a system. Standalone utilities, test files, and pure type definitions are ignored.

File Context & Relationship Manager

Name: File Context & Relationship Manager
Author: murillodutt

bymurillodutt

0•

Herramientas para Desarrolladores

Manages cross-file dependencies by enforcing context headers and ensuring AI reads related system files before modification.

The cellm file-context skill is a specialized utility designed to prevent architectural regressions by enforcing a deep understanding of file relationships. It works by classifying files into categories like Protocol, Shared Data, or Lifecycle systems and injecting standardized 'Related files' headers. This automation ensures that Claude Code follows a 'Read-Before-Edit' rule, forcing the AI to analyze the entire subsystem before making changes to any single component. By mapping non-obvious dependencies and runtime files, it maintains system integrity in complex, multi-file environments like Nuxt 4 stacks.

Características Principales

01Passive detection of multi-file system boundaries based on import patterns

02Enforcement of the 'Read-Before-Edit' protocol to prevent breaking dependencies

03Support for mapping non-obvious runtime dependencies and configuration files

04Automatic classification of file relationships (Protocol, Shared Data, Lifecycle, Pipeline)

05Standardized 'Related files' header generation for system-level documentation

060 GitHub stars

Casos de Uso

01Managing complex state management or database schemas shared across multiple consumers

02Refactoring multi-stage data pipelines where the output of one file is the input of the next

03Documenting and safely editing start/stop/cleanup lifecycle sequences in backend services

Características Principales

01Passive detection of multi-file system boundaries based on import patterns

02Enforcement of the 'Read-Before-Edit' protocol to prevent breaking dependencies

03Support for mapping non-obvious runtime dependencies and configuration files

04Automatic classification of file relationships (Protocol, Shared Data, Lifecycle, Pipeline)

05Standardized 'Related files' header generation for system-level documentation

060 GitHub stars

Casos de Uso

01Managing complex state management or database schemas shared across multiple consumers

02Refactoring multi-stage data pipelines where the output of one file is the input of the next

03Documenting and safely editing start/stop/cleanup lifecycle sequences in backend services