configure-notifications

The skill aims to provide a unified configuration surface for OMX notifications with both native platform support and generic extensibility. The core approach—mutating a local JSON config via jq—is coherent with a configuration-centric capability. However, several security considerations exist: interpolated user inputs into shell-executed jq statements raise command-injection risks if inputs are not properly sanitized; credentials such as webhook URLs, bot tokens, and event templates could be exposed if logs or plaintext config files are not protected; reliance on OpenClaw/clawdbot workflows introduces external process boundaries and potential data leakage if not tightly sandboxed; and there are unverifiable dependencies and binary/tool usage that could act as data exfiltration points. Overall, the footprint is suspicious to high-risk due to potential credential exposure and external tool integrations, and should be treated as suspicious until stronger input sanitization, explicit per-action user prompts, and verifiable, audited dependencies are demonstrated.