Category · Collaboration & tooling

Collaboration & tooling

5 skills Category 4 of 20

This category is about shared human–agent habits: incremental modernization of monoliths, Cursor rules that encode conventions, planned dependency upgrades with breaking changes, static analysis, and lint/format baselines in local dev and CI. Combined with code review and CI/CD, the toolchain becomes a reusable team asset.

In the hub it maps to the “code & collaboration” theme. The five entries match the main hub.

In depth

Legacy modernization

Strangler fig, anti-corruption layers, and feature flags to replace modules incrementally—not big-bang rewrites; emphasize dual-write validation, traffic cutover, and rollback—often read alongside architecture boundaries.

Cursor rules

Author `.cursor/rules` for stack, layout, anti-patterns, and PR self-checks so agents inherit context every session with less repeated explanation.

Dependency upgrades

Separate security patches from major jumps; use changelogs, typecheck, and E2E smoke; lockfiles (npm/pnpm/yarn, Go modules) should be auditable and reproducible.

Static analysis

ESLint, tsc, Semgrep, etc., with CI gates and baselines to pay down debt gradually—balance rules vs. business false positives.

Lint & format

Prettier, EditorConfig, etc., to take formatting out of code review; align hooks and CI with the same rules to avoid noisy diffs.

Back to more skills Skill library (filters)