Komodo

**Quick judgement** Komodo is a powerful, operationally-focused skill for managing infrastructure backed by the Komodo Core platform. It provides broad coverage of server, deployment, stack, build, and procedure operations, making it suitable for teams that already standardize on Komodo for container orchestration and deployments. It is **high-impact but high-risk**: it can start/stop/redeploy production workloads. **What it does well** - Offers a comprehensive set of actions aligned with the Komodo API: listing resources, managing lifecycle (deploy/start/stop/restart), triggering builds and procedures, and accessing logs and stats. - Mirrors the Komodo CLI and API, which should make it straightforward to map existing operational playbooks into AI-driven workflows. - Supports both high-level commands and low-level direct API call patterns (read/execute endpoints), allowing extension to operations not explicitly wrapped. **Key risks / cautions** - **Destructive or highly impactful actions**: stop, restart, deploy, delete-stack, and run-procedure can cause outages or data loss if misused or run on the wrong target. - **Environment-bound access**: requires `KOMODO_ADDRESS`, `KOMODO_API_KEY`, and `KOMODO_API_SECRET`; incorrect scoping or overly-permissive API keys could expose full production control. - **Limited guardrails described**: no built-in confirmation prompts, role scoping, or dry-run mode are evident from the description, so higher-level policies (RBAC, environment separation, human approvals) are recommended. - **External dependency**: relies entirely on the Komodo Core API’s availability and correctness; network or platform issues will directly impact reliability. **Recommended scenarios** - Teams already using Komodo Core for Docker-based infrastructure who want AI assistance for **day-to-day operations, deployments, and incident response**. - On-call and SRE workflows where the assistant needs to quickly **inspect resource state, read logs, and perform controlled restarts or redeployments**. - CI/CD or release processes where the assistant orchestrates **builds, stack deployments, and maintenance procedures** under well-defined guardrails. **When to avoid or restrict** - Environments without Komodo Core, or where Komodo is not the authoritative source of truth for deployments. - Highly regulated or safety-critical production systems unless combined with strict access control, approvals, and limited-scope API keys. - Use cases requiring fine-grained per-action authorization or detailed audit controls beyond what Komodo itself and the surrounding platform provide.