Jo
Languages and libraries
programming language
SHOW HN2026
How HN took it
3 positive, 4 neutral or asking, 1 mixed, 0 critical, out of 8 comments about the project. Thread score +0.29.
Idea: 2 for, 0 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
Coherent design philosophy and timely focus on AI security
From 1 comment
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
aaronon on HN
Capability-based security seen as promising for building AI agents
From 1 comment
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!
deug on HN
Capabilities design is really cool
From 1 comment
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
bon_ai on HN
Objections
Title says compile-time but injection is runtime data; this is enforcement not detection
From 1 comment
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…
hiroto_lemon on HN
Maybe prompt injection guards are unnecessary if security lives outside the model
From 1 comment
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…
benbct on HN
Questions
Why not use API-only agents or database row-level security like pg RLS instead
From 1 comment
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
bon_ai on HN
Alternatives named
API wrappers restricting agent actions
From 1 comment
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…
aaronon on HN
Database-level security such as pg RLS
From 2 comments
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
bon_ai on HNyeah, 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…
aaronon on HN
Runtime sandboxes blocking syscalls
From 1 comment
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…
liu-fengyun on HN