开发者工具2 weeks ago
CrewScore — find missing written guardrails
用23个浏览器本地控件检查AI助手指令的缺失防护措施。
Open source
shmindmaster · HN
CrewScore – coverage check for AI agent prompts (0.6.11)
完整作品展
技术栈
3 projects
用23个浏览器本地控件检查AI助手指令的缺失防护措施。
shmindmaster · HN
CrewScore – coverage check for AI agent prompts (0.6.11)
Mount what your model was never trained on. Unmount it when you're done. Sealed, signed, measured knowledge packs for local models.
@developerpranab · X
YantrikDB: mountable knowledge for local models. A 4B scores 7/53 on the MCP spec published this week; 42/53 with the pack mounted. No retraining.
Knowl is an MCP memory server for Claude Code, Cursor and Codex. What your agents work out survives the session, and a change retires what it replaced — so the answer they read bac
@dat999zx · X
I built Knowl so my AI agents stop forgetting what's true. One memory they all read and write, always current. Runs entirely local — Claude Code, Codex, Cursor. Or for team sync, and reach the same memory from Claude and ChatGPT on the web.