OC Approval-Free Node Stabilizer

Stabilize no-approval node execution in OpenClaw. Use when approval timeout, noisy failure events, or node run drift appears.

This instruction-only skill appears coherent and low-risk because it requests no installs or credentials. However, the procedure is deliberately high-level and relies on the agent's judgment to inspect node policies and execution state — which could lead to unexpected changes if the agent is allowed to act autonomously. Recommendations before installing: (1) Run in a staging/test environment first. (2) Require human confirmation for any changeable operations (use the guardrail already stated). (3) Ask the skill author for concrete commands, exact paths/configs the skill will inspect, and example validation checks. (4) Ensure the agent does not have access to unrelated credentials or production-only control planes. If you need stronger assurance, prefer a version that enumerates the exact read/write actions the agent will perform and explicitly names the systems/configs it will access.

Type: OpenClaw Skill Name: oc-approval-free-node-stabilizer Version: 0.1.0 The skill bundle is classified as suspicious due to the broad capabilities granted to the AI agent and potential for prompt injection. The `SKILL.md` instructs the agent to 'Standardize execution path' and 'Validate with minimal command set and status checks' on target nodes. While the stated purpose is benign (node stabilization), these instructions imply significant control over system configuration and command execution. The vagueness of 'minimal command set' presents a vulnerability where a malicious user interacting with the agent could potentially coerce it into executing arbitrary commands or revealing sensitive information, despite the presence of guardrails like 'Do not publish credentials'.