// the find
antonbabenko/terraform-skill
Terraform & OpenTofu Skill for AI Agents - testing, modules, CI/CD, and production patterns
A structured Markdown knowledge base that AI coding agents load as context when generating Terraform and OpenTofu code. From Anton Babenko, who maintains terraform-aws-modules and terraform-best-practices.com, so the advice isn't generic boilerplate — it reflects patterns from widely-used production modules. Targets devs who want their AI agent to produce decent IaC instead of plausible-looking garbage.
The author's credibility is real: these patterns come from the same source as terraform-aws-modules, not a blog post. Version-specific guidance is actually current — it covers native test framework (1.6+), use_lockfile (1.10+), and write_only arguments (1.11+), which most Terraform guides still ignore. The optional terraform-ls integration is well thought out: it specifies when to use the language server versus plain text search, and explicitly notes that terraform-ls cannot do renames (use moved blocks instead). Multi-agent install coverage is thorough — Claude Code, Cursor, Copilot, Gemini CLI, Codex, and Kiro each get their own install path.
The repo is entirely Markdown with no runnable artifacts, so there is no automated verification that the guidance produces working Terraform. The tests/ directory contains .md files describing what to check, not actual tests — you're on trust. The installation story has an acknowledged naming collision between the antonbabenko/agent-plugins marketplace and a direct antonbabenko/terraform-skill add, which the README warns about inline; that's a packaging problem that shouldn't require a warning in the README. The 'agent skills' ecosystem this bets on is not standardized — each agent has a different discovery path and the npx skills add shortcut relies on agentskills.io, a third-party registry with no obvious governance. If that ecosystem fragments or stalls, this format becomes just a CLAUDE.md you cloned manually.