A clearer view of work done with agents.
When several agents work for you at once, it can be hard to see what actually got done. I built this work log for that problem and use it in my own projects.
- used in my own projects
- In use
- implementation carried out by agents
- Agent built
- code and interface checks kept with the project
- Recorded checks
Based on project records and recorded checks. These describe the work, not product adoption.
The job
I wanted to make sense of completed work without reading every conversation. The aim was a useful record that I could come back to.
What the agents built
Agents built the product and checked the implementation. Their work covered the software behind it as well as the interface.
What I decided
I chose the problem and set the requirements. The agents made implementation choices within that brief, with changes recorded and reviewed.
How it was checked
The project records include code and browser checks. This account draws on those records. Those earlier checks were not rerun for this shorter write-up.
What this doesn’t show
The interface detail uses demonstration data, not a real account.
These checks do not show whether other people find the product useful or keep using it.