L4 Acting in the world
Extension hooks
Fixed points in the loop where plugins can watch, change or veto what the agent does.
Why it matters
Hooks let a team add policy, logging and tools without forking the harness. They also run with the agent's full authority, so their order, failures and changes must be predictable and visible.
What to build
Ordered command hooks, paired veto, trusted hashes
The default the research recommends. The reasoning and the code behind it are in the app.
What it takes, by kind of agent
- Conversational assistantRecommended
- CodingRecommended
- Deep researchOptional
- Browser/GUIOptional
How well it is studied
- Harnesses checked
- 15
- Checked facts behind it
- 164
See the 164 checked facts and how each harness does it
The facts, the code lines they come from and the rules for this topic are in the app. Sign in with Google; new accounts are approved by hand.
Also in L4 Acting in the world
- Tool dispatchSettled practice
Defines each tool once, checks every call before it runs, and hands every failure back to the model as a result it can fix.
- Assistant Must-have
- Coding Must-have
- Research Must-have
- Browser Must-have
- Perception and groundingDepends on agent type
Lets a browser or screen agent act on named elements, look again before every move, report missed actions plainly, and keep screenshots from flooding the context.
- Assistant Not applicable
- Coding Not applicable
- Research Not applicable
- Browser Must-have