en un clic
better-code
better-code contient 7 skills collectées depuis telekom, avec une couverture métier par dépôt et des pages de détail sur le site.
Skills dans ce dépôt
Design the target architecture for a decomposed feature. Produces service blueprint, spec-to-target mapping, project scaffold, ADRs (Architecture Decision Records), and C4 diagrams. Use after 'decompose' when you have spec.json and need to design how to build it in the target platform.
Assess a discovered codebase and produce a feature-based migration plan. Decomposes the system into business features, classifies each with a migration strategy, maps to target architecture, and sequences migration waves. Requires discovery output in .migration/discovery/. Use when the user says 'plan the migration', 'assess for migration', or 'what's the migration strategy'.
Decompose a selected feature into a complete behavioral specification for migration. Extracts every business rule, data model, sequence flow, error path, and edge case — structured as JSON for the transform step. Also generates Mermaid/PlantUML diagrams. Use after 'assess' when the user selects a feature to migrate.
Discover and document a legacy codebase for migration. Scans source code, builds a knowledge graph via graphify, and produces structured documentation in .migration/. Use when the user wants to understand, map, or prepare a codebase for modernization — even if they just say 'analyze this project' or 'what does this code do' in the context of migration work.
Generate target code from architecture specifications using GitHub's spec-kit methodology. Follows all 6 spec-kit phases: constitution, specify, clarify, plan, tasks, implement. Produces compilable, tested code with full traceability to behavioral spec. Use after 'architect' when you have blueprint.json + mapping.json and want to generate the target project.
Install and validate the one-modernizer environment. Creates venv, installs the CLI, detects host/OS. Run on first use or when something seems broken.
Post-implementation quality gates with self-healing fix loop. Verifies generated code compiles, tests pass, APIs respond, contracts are complete, and spec coverage is 100%. Automatically fixes issues and re-verifies until all checks pass or no progress is made. Run after 'implement'.