Sre Engineer

Use when defining SLIs/SLOs, managing error budgets, or building reliable systems at scale. Invoke for incident management, chaos engineering, toil reduction, capacity planning.

This skill is essentially a structured SRE knowledge and template pack. Before relying on it: (1) review any generated automation, Terraform, or shell scripts carefully — do not run them without inspection and appropriate least-privilege credentials; (2) verify monitoring/alert configs fit your environment and secrets are not embedded in generated outputs; (3) treat runbooks and incident actions as drafts that require human validation; and (4) consider auditing any IaC with your standard security/CI checks before applying to production.

Type: OpenClaw Skill Name: sre-engineer Version: 0.1.0 The skill bundle is classified as suspicious due to the presence of risky capabilities, specifically the use of `subprocess.run` with `shell=True` for `kubectl` commands in `references/automation-toil.md`, which poses a command injection risk if the command string is derived from untrusted input. Additionally, `references/incident-chaos.md` demonstrates the execution of highly privileged and disruptive system commands (`tc`, `iptables`, `kubectl delete`) via `subprocess.run` for chaos engineering purposes. While these actions are plausibly needed for the stated SRE automation and chaos engineering objectives, they represent significant security risks if misused or executed in an uncontrolled environment, lacking clear malicious intent but demonstrating broad, powerful permissions.