Public-safe Mac multi-instance deployment skill for generic OpenClaw-style workspace layout, boundary notes, sample config, and pack validation.

This appears to be a safe, documentation-only template. Before you use it: (1) verify any scripts the checklist tells you to run (./validate_repo.sh, ./generate_public_pack.sh) actually exist in your repo and inspect them before execution — the skill doesn't include those scripts; (2) copy the example env to a private path as instructed and never commit real secrets to the public repo; (3) note that references/instances.md mentioned in instructions is not present in the skill files listed here, so confirm you have the expected reference docs; (4) if you allow agents to invoke skills implicitly, be aware this skill's agent metadata permits implicit invocation — that's normal for helpers but ensure your agent policies match your risk tolerance.

Type: OpenClaw Skill Name: mac-multi-instance-deployment Version: 1.0.4 The skill bundle is a template for organizing Mac-based multi-instance workspaces in OpenClaw, focusing on separating public and private data. The instructions in SKILL.md and the examples (e.g., setup-example.md, quickstart.md) promote security best practices by guiding the user to keep secrets in a 'private' directory and providing validation steps to prevent accidental data leakage. No malicious execution patterns, data exfiltration, or harmful prompt injections were identified.