Public framing
The interesting problem is not clicking buttons. It is adapting to messy workflows, grounding actions in real user context, learning from failure, and knowing when to ask the human.
This project explores a browser agent that can help with high-friction application flows without turning into a mass auto-apply tool.
What it is for
- Understanding recurring web workflow patterns.
- Drafting form answers from approved source material.
- Remembering where workflows break.
- Asking for human review before any sensitive or final action.
What it is not
- Not a mass application bot.
- Not a recruiter bypass system.
- Not a system that fabricates qualifications.
- Not something that submits without explicit approval.
Questions
- How should the agent represent form state and user intent?
- What failures should be remembered for the next attempt?
- Where should the system pause and ask Esther instead of guessing?