How your agents are doing, and what to change
Agents rarely fail loudly. See the pattern, change one thing.
A retry here, a lint error there. Health reduces a date range to one sentence about the fleet and the agents that moved it, names the friction you can remove today, and turns patterns into recommendations with the evidence attached.
One sentence about the fleet
One strip says what this window looks like, what waits on you and what you changed. Below it, the agents that moved the number, and the friction you can remove today in one click.
- Per-agent scorecards against the previous window
- Errors split into behaviour, harness and infra
- Estimates checked against actual hours, per task
What cost you time
Which tools you keep denying, each one click from a rule. Repeated errors, and how long agents waited on you. Lessons from retros, grouped by how often they repeat.
- A denied tool is one click from an auto-approve rule
- A triage run mutes noise and files a fix task
- Approval wait time, per session
Know the cost before the bill does
What you spent, what drove it, and where the month is heading. Split it by project, model or agent, and export it as CSV. Productivity and autonomy sit beside the spend.
- Cost and tokens against the live quota window
- A projected monthly figure
- The Advisor turns patterns into recommendations
screenshots
The health pages, up close.
Click any screenshot to open it full size. The projects are fictional.
Run your agents like a fleet.
Free on GitHub. One binary on your machine, no cloud, no account. Your sessions show up the moment it starts.