loop-verifier
The independent gate for your agent loop or swarm. Use when you run an autonomous or scheduled agent loop OR a cheap-volume / open-weight / open-source agent swarm (Claude Code /loop, scheduled tasks, hooks, Codex Automations, CI-triggered agents, cron'd agents, open-source agent frameworks like CrewAI / LangGraph / AutoGen, hundreds of parallel sub-agents) and need an objective verifier that is NOT the agent that did the work — before the loop merges, deploys, trades, pays, publishes, or SAVES A RESULT AS A REUSABLE SKILL. Turns an independent judgment verdict into an exit code your loop can gate on, plus a portable signed proof attached to whatever the loop ships or saves — so a confident-but-wrong output never gets kept as a skill and replayed forever. The engine can be the cheapest model you like; the gate has to be one that isn't the engine. Keywords - loop, gate, verifier, swarm, sub-agent, parallel agents, skill, save skill, open-weight, open-source, open model, cheap model, CrewAI, LangGraph, AutoGen,
来源信息
- 仓库
- babyblueviper1/invinoveritas
- 最近来源活动
- 2026年7月29日 10:36
- 检测到的 SKILL.md 语言
- 英语
- 星标
- 5
- 分支
- 0
安装方式
默认使用会先检查来源的 Prompt;你也可以切换为直接命令,或下载本地副本。
检查来源文件
决定是否安装前,请先阅读 SKILL.md,以及 SkillsMP 当前展示的配套文件。