How do I catch an agent that claims a task is done when it isn't?

Grade the state a session actually left behind, not the agent’s closing message. A completion grader compares an end state, a payment reference, a ticket status, a calendar entry, against what the user asked for, so the words “done” or “confirmed” in a transcript carry no weight by themselves.

ReliabilityBench shows why the distinction matters. Its travel domain, built on tools like search_flights, hold_flight, and confirm_booking, scores 87.5% pass@1, and the paper’s own failure analysis names the gap directly: agents fail to provide payment information for booking confirmation, a missing-context problem, not a broken call. Picture what that looks like in a trace: every step up to confirm_booking can return a valid response, so a checker that only reads step outcomes passes the run, even though confirm_booking never fires and no reservation exists.

The fix is to check the system the agent was supposed to change, not the trace of calls that got it there. If the booking system shows no reservation, the run failed, whatever the last message claimed.

sources

keep reading

More on this.

How do you find a failure mode nobody defined a check for?By clustering traces on how the failure behaves instead of grading against a named check, since a check only ever catches a failure someone already named.answer →What causes a retry storm in an agent?A permanent error the retry policy treats as temporary, most often a tool name the model invented; one ReAct benchmark saw 466 of 513 retries hit exactly that.answer →What happens when an agent's conversation exceeds its context window?Depends on the provider: some APIs reject the request with a hard error, others silently drop the oldest turns and answer as if nothing was cut.answer →Everything on failure modesThe definition, and every question answered under failure modes.concept →

Two ways to run Tessary.

Tessary is an open-source agent reliability platform. Cloud and self-hosted run the same workflow on the OpenTelemetry traces your agent already emits.

Tessary Cloud

We host it for you. Send your first trace with nothing to deploy and no model key.

what's includedper organization
traces
10,000 per calendar month
stored trace data
1 GB
retention
30 days
model credit
$10, one-time, for triage and root-cause analysis
credit card
not required
start for free

Self-hosted Tessary

Run the open-source code on your own infrastructure with one command. Add your own model key for triage and root-cause analysis.

Self-host Tessary for me by following https://github.com/tessaryai/tessary/blob/main/setup.md

docker compose -f oci://docker.io/tessaryai/tessary:compose up -d -y