Back to Registry
aiv.shell.state_probe
AGENTIC
Executes a sandboxed read-only shell command and compares stdout to an expected value, probing whether an agent action actually changed system state.
Scorecard
| Determinism | deterministic |
| Evidence Quality | hard-state |
| Intended Use | eval-and-train |
| Gating Required | No |
| Permissions | subprocess:readonly |
Attack Surface
| injection risk | low |
| format gaming risk | low |
| tool spoofing risk | low |
Test Fixtures
9 total
| Type | Count |
|---|---|
| Positive | 3 |
| Negative | 3 |
| Adversarial | 3 |
Metadata
| Version | 0.1.0 |
| Domain | aiv |
| Task Type | shell_state_verification |
| Contributor | vr.dev |
| Source | arXiv:2602.00575 |
Use in SDK
# CLI
vr verify --verifier aiv.shell.state_probe --ground-truth '{"order_id": "ORD-42"}'
# Python
from vrdev import verify
result = verify("aiv.shell.state_probe", ground_truth={"order_id": "ORD-42"})
# API
curl -X POST https://api.vr.dev/v1/verify \
-H "X-API-Key: vr_live_..." \
-d '{"verifier": "aiv.shell.state_probe", "ground_truth": {"order_id": "ORD-42"}}'