Have something to say?
Flagged questions from agents, with a filter so I can answer them without parsing the whole conversation
When I work with a team of agents there is constant back and forth. If they need input from me they ask the question, and it quickly gets lost in the noise of the conversation. Later they refer back to it vaguely, for example "still waiting on an answer from Matt for my form submission question." By then I have to reconstruct what they actually asked. It would help if agents could flag a question as waiting on me, and I had a filter that showed only those flagged questions so I could read and respond to them in one place, without parsing everything.
Team creation: short names, a recycle button, and regional name bias
When I create a team of agents, Kosmos automatically populates names. I just stood up two teams, and this is what it chose: Quality Assurance and Testing - Chamari, QA Lead - Adewale, Functional Test Specialist - Delroy, Accessibility and Usability Tester - Jasleen, Test Automation Engineer Cybersecurity - Caio, Security Lead - Naoko, Security Configuration Analyst - Anirban, Identity and Access Specialist - Revathi, Incident Preparedness Specialist - Kamau, Security Awareness Writer Auto-naming is useful. Three changes would make it much easier to live with as the roster grows. 1. Short name toggle A toggle at team creation that gives every agent a non-colliding short name, 3 or 4 characters, easy to pronounce and easy to type. These names get used in kosmos msg, in @-mentions, and on the board. They should stay unique on the machine. 2. Recycle button for one agent's name Next to an individual generated name, a control that rolls only that name to a new candidate and leaves the rest of the team as they are. 3. Regional name bias Optional, based on where the user is. If they are in India, prefer Indian names. If they are in Germany, prefer German names. Same idea for other regions. Today's mix is global; a local default would feel more like a team they actually hired.