[{"data":1,"prerenderedAt":183},["ShallowReactive",2],{"project-48412229":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":10,"maker":11,"posted_at":12,"year":13,"points":14,"comments":15,"rank_in_year":16,"shown_in_year":17,"top":18,"tier":19,"category":20,"subcategory":21,"format":22,"keywords":23,"reception":29,"sentiment":30,"stance_mix":31,"scored_comments":36,"aspects":37,"strip":41,"post_text":43,"what":44,"said":59,"maker_in_thread":119,"shape":121,"grid":124,"attempts":146,"same_week":147,"maker_shows":33,"maker_hidden":175,"reader_version":176,"card":177},"48412229","Jo","AI-native language to catch prompt injection at compile-time","title","Show HN: Jo – AI-native language to catch prompt injection at compile-time","https:\u002F\u002Fgithub.com\u002Ftypescope\u002Fjo","github.com","kiru_io",1780665951,2026,10,9,3079,4366,71,"shown","languages","programming language","other",[24,25,26,27,28],"prompt injection","capability security","typed language","sandboxing","ai agents","curious",0.286,{"positive":32,"neutral":33,"mixed":34,"critical":35},3,4,1,0,8,{"idea":38,"docs":40},{"positive":39,"critical":35},2,{"positive":35,"critical":35},[32,32,32,32,42,42,42,32,35],5,null,{"problem":45,"audience":43,"pricing":43,"motivation":53,"time_to_build":43,"stack":57,"openness":58,"stage":58,"team_size":58},{"value":46,"source":47,"evidence":48},"Limiting the damage prompt injection can cause by enforcing capabilities at the language level","commenter",[49],{"id":50,"by":51,"text":52},"48413287","liu-fengyun","That is true -- the title does not precisely state what Jo is. It bounds the dangers that prompt injection can cause to arbitrary granularity. I (the author of Jo) built Jo…",{"value":54,"source":47,"evidence":55},"Maker asked whether sandboxing could be solved at the language level, aware of business logic",[56],{"id":50,"by":51,"text":52},[],"unknown",{"praise":60,"objections":82,"requests":97,"questions":98,"alternatives":103},[61,68,75],{"value":62,"source":47,"evidence":63},"Coherent design philosophy and timely focus on AI security",[64],{"id":65,"by":66,"text":67},"48420457","aaronon","I like that this feels driven by a coherent design philosophy. and the emphasis on AI security is very timely as AI-generated code becomes more common",{"value":69,"source":47,"evidence":70},"Capability-based security seen as promising for building AI agents",[71],{"id":72,"by":73,"text":74},"48416775","deug","I am building an AI agent using Jo. I believe Jo has great potentials in this field with its capability-based security features. Great work!",{"value":76,"source":47,"evidence":77},"Capabilities design is really cool",[78],{"id":79,"by":80,"text":81},"48420688","bon_ai","the capabilities design is really cool, however to protect against prompt injection to unauthorized db access, couldn't we just use api only agent or db features like pg RLS",[83,90],{"value":84,"source":47,"evidence":85},"Title says compile-time but injection is runtime data; this is enforcement not detection",[86],{"id":87,"by":88,"text":89},"48412607","hiroto_lemon","Injection is runtime data, so 'compile-time' overstates it. A type system can taint-track — mark untrusted input, block it from a privileged sink. Valuable, but that's…",{"value":91,"source":47,"evidence":92},"Maybe prompt injection guards are unnecessary if security lives outside the model",[93],{"id":94,"by":95,"text":96},"48457635","benbct","Nicely done. This isn't a criticism, but i wonder if a well designed system needs prompt injection guards at all. Provided all security happens outside of the models and models…",[],[99],{"value":100,"source":47,"evidence":101},"Why not use API-only agents or database row-level security like pg RLS instead",[102],{"id":79,"by":80,"text":81},[104,110,115],{"value":105,"source":47,"evidence":106},"API wrappers restricting agent actions",[107],{"id":108,"by":66,"text":109},"48420929","yeah, not really, actually. in my opinion, they may work, but with tradeoffs api wrappers are safe, but they kill flexibility of AI agent and will have massive maintenance…",{"value":111,"source":47,"evidence":112},"Database-level security such as pg RLS",[113,114],{"id":79,"by":80,"text":81},{"id":108,"by":66,"text":109},{"value":116,"source":47,"evidence":117},"Runtime sandboxes blocking syscalls",[118],{"id":50,"by":51,"text":52},{"comments":35,"answered":35,"key":120},[],{"commenters":36,"max_depth":39,"maker_replies":35,"first_comment_mins":122,"lifespan_hours":123},28,90,[125,127,129,131,135,137,138,142,144],[94,35,32,126,95,35],"project\u002Fneutral\u002Fidea",[87,34,32,128,88,35],"project\u002Fmixed\u002Fidea",[50,34,32,130,51,35],"maker\u002Fneutral\u002Fidea",[132,39,32,133,134,35],"48413261","project\u002Fneutral\u002Fdocs","tristones1983",[65,32,42,136,66,35],"project\u002Fpositive\u002Fidea",[72,33,42,136,73,35],[139,42,42,140,141,35],"48415382","maker\u002Fpositive\u002Fnone","yinlin_hu",[79,143,32,126,80,35],6,[108,143,35,145,66,35],"commenter\u002Fneutral\u002Fidea",[],[148,153,158,162,166,170],{"id":149,"name":150,"points":151,"category":152},"48445554","Performative-UI",1181,"dev_tools",{"id":154,"name":155,"points":156,"category":157},"48433756","Lathe",402,"productivity",{"id":159,"name":160,"points":161,"category":152},"48408854","Absurdly Optimized Pancakes",340,{"id":163,"name":164,"points":165,"category":152},"48447806","Gitdot",334,{"id":167,"name":168,"points":169,"category":152},"48478469","Extend UI",252,{"id":171,"name":172,"points":173,"category":174},"48396004","Uruky",236,"web_saas",false,"v2",{"id":178,"name":5,"line":6,"ai":175,"cat":20,"year":13,"points":14,"comments":15,"rec":29,"s":179,"top":18,"kw":180,"maker":11,"domain":10,"tier":19,"attempts":35,"strip":181,"url":9,"mix":182,"praise":62,"objection":84,"sub":21},48412229,0.29,[24,25,26,27,28],"333355530",[32,33,34,35],1793991156061]