agent runs
Traces
Every run, end to end: the agent's turns and tool calls, the verifier's functional / secure verdict, and the files left in the workspace. F∧S = solved, (1,0) = tests pass but the vulnerability remains.
…
runs
0
solved (F∧S)
0
func failed (F✗)
0
vuln left (F✓S✗)
0
models
—
solve rate
loading…
Tier:
Outcome:
Filter:Group:Sort:loading…