A collection of skills for AI agents, installable via vercel-labs/skills.
Go coding best practices: OOP with interfaces and pointer receivers, control flow patterns, and Go proverbs. Includes reference docs on the options pattern, typed nil in interfaces, and test error assertions.
Plans a Go project package structure refactor following Screaming Architecture and the Dependency Rule. Produces a comprehensive plan document and a frozen go-arch-lint config that serves as the immutable success criterion. Triggers when a project has layer-named packages (usecases/, repository/, domain/, etc.) that should be renamed to reflect domain concepts.
Profiles Go build performance using -debug-actiongraph and the actiongraph CLI. Identifies the slowest packages, aggregates compile time by namespace, and traces why an unexpected package is being compiled.
Improves test coverage by grouping gaps by package and delegating each to a focused subagent. Subagents are permitted to refactor for testability, not just add tests. Supports an optional path argument for monorepos. Retries up to 3 rounds before surfacing remaining gaps to the user.
Fixes golangci-lint violations by grouping them by rule and delegating each group to a focused subagent. Avoids //nolint suppression in favor of proper structural fixes.
Guidance for embedding files in a Go binary, covering both simple and overlayed filesystem approaches.
Install all skills:
npx skills add brahmlower/skillsInstall a specific skill:
npx skills add brahmlower/skills --skill go-improve-coverageInstall globally (available across all projects):
npx skills add brahmlower/skills -gInstall to a specific agent (e.g. Claude Code):
npx skills add brahmlower/skills -a claude-code