Open Agent Observatory · Public experiment receipt
crewai / tool-failure
FRAMEWORK_RUNTIME_REPRODUCED. This is framework-runtime reproduction on a deterministic public/synthetic task, not model-performance validation or project-wide verification.
01 / Identity
Exact bounded claim.
- Receipt SHA-256
sha256:14b921e9c9d5f02420996b46dc64b0158b34006b7711f3f99078389d237dce2b- Origin
OPEN_AGENT_LAB_PUBLIC_SYNTHETIC- Upstream commit
92eb5f91830cde5a7c4be921cb295c5523649d13- Licence
- MIT
- Suite / partition
open-agent-runtime-v1/ development- Environment
sha256:fd7cc118014017039a60c11f25d6c6ea30d42abfa8857e4e8b378ebfe4958af3- Runtime model
- framework-native-deterministic / no-network-mock / 1.15.18
- Outcome / label
- FAILURE / TOOL_FAILURE
- Started / completed
- 2026-09-04T00:50:54.433578Z
2026-09-04T00:50:54.457294Z
02 / Normalized events
Metadata only.
Raw prompts and outputs were not retained. Events expose bounded action classes, permission decisions, timing, retries, and supported error labels.
| # | Action | Target | Permission | Retry | Error | Latency ms | Warning |
|---|---|---|---|---|---|---|---|
| 0 | task | synthetic-agent-0 | synthetic-local | 0 | TOOL_FAILURE | 22.693 | WARN |