Terraform Ai Skills

Use when bulk-managing Terraform modules at scale — upgrading providers across AWS, GCP, Azure, or DigitalOcean repositories, standardizing GitHub Actions wo...

This skill appears to do what it claims, but it executes shell scripts that clone, modify, and push to many repositories — potentially at scale. Before installing or running: 1) Review the scripts (scripts/*.sh and run-with-provider.sh) line-by-line in a safe environment. 2) Test on a single non-production repository (the README and SKILL.md explicitly advise this). 3) Provide GitHub credentials with the minimum necessary scopes (use a fine‑grained token or repository-scoped PAT; the docs recommend contents:write, workflows:write, pull-requests:write only if needed). 4) Use DRY_RUN or create PRs (CREATE_PR=true) rather than direct commits while validating behavior. 5) Avoid exposing other secrets in config files; use GitHub secrets when integration is required. 6) Rotate and limit tokens after use and keep audit logs of operations. If you need higher assurance, have a trusted engineer audit the scripts for any unexpected network calls or external endpoints before running at scale.

Type: OpenClaw Skill Name: terraform-ai-skills Version: 0.0.2 The skill bundle is designed for legitimate Terraform module management, utilizing standard DevOps tools like `terraform`, `git`, and `gh`. It requests broad `filesystem` and `network` permissions, which are plausible for its stated purpose. However, the `run-with-provider.sh` script executes `bash "$SCRIPT_PATH" $ARGS` where `ARGS` is passed directly from user input without explicit sanitization. This creates a shell injection vulnerability (potential RCE) if a malicious user or a compromised AI agent provides crafted input for `ARGS`. While the skill's internal prompts and documentation emphasize safety and do not demonstrate malicious intent, this critical vulnerability, combined with the broad permissions, classifies it as 'suspicious' rather than 'benign' or 'malicious'.