Meditate

What to consider before installing: - This skill will read your conversation history and create/update files in ~/meditate/ (profile.md, topics.md, insights.md, feedback.md, archive/). The registry metadata did not declare those config paths — confirm you are comfortable with local storage in your home directory. - The skill states it will never send data externally, use credentials, or perform actions, and has strict sandbox rules in the docs, but those are behavioral constraints the agent must obey; the code is not present to enforce them. If you use this skill, periodically inspect ~/meditate/ contents and permissions. - If stored insights may contain sensitive details, consider restricting file permissions (chmod 700), placing the directory on encrypted storage, or running the skill in a segregated environment/account. - The template shows a shell command for initial setup (mkdir -p ...). Decide whether you will run that manually (safer) rather than allowing automated file creation. - Test the skill with non-sensitive topics first to verify it behaves as promised. If you need stronger guarantees (encryption, explicit consent prompts, platform-enforced sandboxing), request those features or avoid installing until they are provided. Confidence note: assessment is medium because the skill is instruction-only (no code to analyze) and the main concerns are transparency about local file I/O and privacy safeguards rather than evidence of malicious intent.