external agent baseline

SOLARI LIVE

Example Repository Agent · Complete Demo Checkout · run a6bdb75 · 3d ago

Reliability

0%0 / 1 runs passed

95% confidence interval 0.0% – 79.3% over 1 scored run

Runs

1 scored

Failed

no infra errors

p95 duration

5.7s

p50 5.7s

Avg steps

—

avg 5.7s

Run-to-run flips

how often identical repetitions disagreed

Reliability by perturbation

Worst first. The whisker is the 95% Wilson interval — small samples are wide on purpose.

Baseline

0%0 / 1

Baseline

0%0 / 1

Only scored failures. Infrastructure errors are excluded.

Agent error

1100%

Every cell is one run. Select one to open its evidence.

Outcome of every run, by perturbation and repetition. Select a cell to open that run.
Perturbation	Run 1	Rate
Baseline	Baseline, run 1: Failed	0%

Failures grouped by category and cause, so a repeated defect reads as one problem.

Agent error
browserType.connectOverCDP: WebSocket error: connect ECONNREFUSED 127.0.0.1:44045 Call log: - <ws connecting> [redacted-session-endpoint] - <ws error> [redacted-session-endpoint] error connect ECONNREFUSED 127.0.0.1:44045 - <ws connect error> [redacted-session-endpoint] connect ECONNREFUSED 127.0.0.1:44045 - <ws disconnected> [redacted-session-endpoint] code=1006 reason=
1 run · 59aca6b

Open one to see the assertion that failed and the trace that led there.