Start
always-on-agent-safety
always-on-agent-safety - Skill Dossier
always-on-agent-safety
Practical safety guidance for persistent agents, covering privacy, cost, scope boundaries, psychological risks, and security controls. Use when reviewing an always-on agent before deployment or expansion. NOT for storage architecture, input-feed design, or abstract alignment theory.
Uncategorized
Allowed Tools
ReadWriteEditBashGlobGrepTaskWebSearchWebFetch
Skills use the open SKILL.md standard — the same file works across all platforms.
Install all 544 skills as a plugin
claude plugin marketplace add curiositech/windags-skills
claude plugin install windags-skills
Claude activates always-on-agent-safety automatically when your task matches its description.