agentsclimarketplace

Project evolver mapper

Skill Vit0lze/Project-Evolver-Framework/Skill Files/project-evolver-mapper

Framework de agentes de IA para mapear e corrigir arquitetura de software. / AI-agent framework to map and heal software architecture.

Install
npx -y skills add Vit0lze/Project-Evolver-Framework --skill project-evolver-mapper

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Systematic codebase mapping for evolution using an incremental 'Bite and Blow' strategy. Use when the user wants to understand, refactor, or evolve a complex codebase that exceeds context limits, or when requested to map a project for future work.

SKILL.md

2.4 KB, as published. Nobody here has run it

Project Evolver Mapper

This skill enables systematic mapping of complex codebases to prepare for future evolution. It uses an incremental strategy to digest large projects and produces a high-density global map.

1. Proactive Vision Interview (The "Why")

Goal: Deduce or extract the 'Project Purpose' before mapping.

  1. Deduce: Briefly scan root files (README, package.json, main entry) to guess the project's core purpose.
  2. Ask: Present this deduction to the user.

    "I see this appears to be a [Project Type] aiming to [Goal]. Is this accurate? What is your primary evolution goal (e.g., refactor, feature add, rewrite)?"

  3. Validate: Do not proceed to deep mapping until the user confirms the vision.

2. Chained Mapping Workflow

Execute this loop to map the codebase.

Phase A: Scan & Strategy

  1. List all files.
  2. Identify large files (>30k tokens) that need splitting.
  3. Plan the mapping order (typically Core -> Utils -> Features).

Phase B: Chunk & Draft ("Bite")

For each file (or chunk):

  1. Split using scripts/chunk_reader.py if the file is too large. 
    python scripts/chunk_reader.py path/to/large_file.py --size 30000
  2. Analyze the chunk. Load references/mapping-patterns.md for templates.
  3. Draft a momentary map in drafts/.
    • Create drafts/MAP_[filename].md.
    • Extract: Logic Anchors, Dependencies, Breaking Points.

Phase C: Synthesize ("Blow")

  1. Combine all draft maps into a single GLOBAL_MAP.md in the root (or brain/ folder if in agentic mode).
  2. Format: strict Markdown.
  3. Language: All output MUST be in English.

3. Output Requirements

  • Drafts: Temporary, verbose analysis stored in drafts/.
  • GLOBAL_MAP.md: The final deliverable. High density, low fluff.
  • Glossary: Include a glossary of domain terms found in the code.

4. Constraint Checklist

  • "Context Safety First": Never read entire large repos at once.
  • "English Only": Analysis must be in English.
  • "Evolution Ready": Focus on changeability (coupling, cohesion), not just description.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.