Use Honeycomb with AI agents, safely
Teaches an agent to confirm the team, environment and dataset before querying, choose the right read tool, keep results small, and ask before any change. It does not modify Honeycomb without your approval.
Use it when
- You are connecting Honeycomb to an AI agent for the first time.
- You want to know which environment and dataset an agent is querying.
- You need to stop an agent from making changes to boards, triggers or SLOs without asking.
Tools it uses
get_team_infolist_environmentslist_datasetsget_environmentsearch_columnsget_dataset_columnsget_tracerun_unsaved_queryget_query_historyget_slosget_triggersget_datasetfeedback
Before you investigate anything, check which Honeycomb environment and dataset you are looking at, and tell me what you can see.
You are connected to team Acme, with 2 environments: production and staging. Your api-gateway data is in the production environment, dataset api-gateway, which has 14 columns including duration_ms and status_code. I have not changed anything. Tell me the time window and I will run a read-only query.