Avdi Grimm avatar

Avdi Grimm

Community

@avdi · St. Louis, USA

1,838Followers
|
213Public Repos
|
8Published Skills

Avdi Grimm maintains skills for Claude Code environments: git worktrees, subagent definitions, code knowledge graphs, devcontainers, and agent code-review frameworks.

Skills Distribution
DomainDeveloper To...Agent Configuratio.. (30%)Devcontainers & Cl.. (25%)Code Review & Audi.. (20%)Git Worktree & Dot.. (15%)

Agent Skills by Avdi Grimm

Showing 8 vetted skills indexed across 1 GitHub repositories.

Frequently Asked Questions About Avdi Grimm

FAQPage Schema
What tasks can I accomplish using Avdi Grimm's skills?

You can run GitHub Project board upkeep with orphan PR detection, manage dotfiles via chezmoi, query a codebase knowledge graph with Cypher for callers and dead code, operate git worktrees safely, define Claude Code subagents with MCP servers, port agent guidance between repos, build auditor review frameworks, and make devcontainers runnable on cloud agent platforms.

Who are these skills designed for?

Developers working inside Claude Code and similar agent environments: maintainers of devcontainer-based projects, engineers writing or debugging subagent definitions, teams bootstrapping agent-driven code review, and anyone managing dotfiles, git worktrees, or cross-repo agent guidance ports.

How do I use the clouddev skill in practice?

Load it when porting a project to a cloud agent platform. It guides the platform tier model, the prepare/boot phase split, authoring the clouddev.yml manifest with entrypoint scripts, path mapping, publishing a dev image to a registry, and splitting docker-compose so backing services start without the app container.

When should I load the worktree skill?

Load it before any git worktree operation in Claude Code: creating, entering, resuming, listing, removing, or troubleshooting. It covers entering and re-rooting sessions, which wt commands are safe to shell out to, and recovering a worktree whose provisioning hooks failed partway through setup.

What does the defining-subagents skill fix?

It fixes subagents that cannot see a tool they should have. It documents .claude/agents/*.md frontmatter fields, what tools a dispatched subagent actually receives, and the mcpServers frontmatter declaration that works around the Agent-tool MCP-inheritance gap when wiring servers like playwright, ripgrep, or github.