Agent detail

Sandbox Runner

Executes agent-generated code in isolated sandboxes with resource limits and network controls.

Audit-first trust modelProvenance awareEnterprise policy controls

VerifiedPassedLow Risk88ScoreEnterprise SafeSource VerifiedSecurity FailedScanned 29 Mar 2026

Description

What this agent does and how it is scoped.

Executes agent-generated code in isolated sandboxes with resource limits and network controls. Published and verified by TrustAgent.

audit-firstenterprise-readyexecutionisolationorchestrationsandboxtrustagent-published

Trust Snapshot

Trust badges

Granular trust badges with evidence-driven status.

Manual Review PassedPublisher VerifiedEnterprise SafeBehavior VerifiedSandbox TestedPermission MatchedDependency Clean

Provenance

Source type

GITHUB

Publisher

TrustAgent

Source license

Apache-2.0

Commit hash

d6e0f4a8b2c7d1e5f9a3b7c1d5e9f3a7b1c5d9e3

Version hash

adae64b078b463f222d1c3ee5bae21b95c00a78bfadf3551a776ef4dad01432d

Audit Evidence

Version History

Hash-locked versions, source metadata, and drift context.

Version	Commit	Hash	Source	Created
1.0.0	d6e0f4a8b2c7d1e5f9a3b7c1d5e9f3a7b1c5d9e3	adae64b078b463f2	TrustAgent	3/29/2026, 8:34:12 PM

Reviews

Brianna BuyerVerified install

Great trust signals and clear audit evidence.

Public trust report

Sandbox Runner completed the Trust Agent automated review pipeline for agent listings. Stage 1 configuration integrity scored 84/100 and stage 2 behaviour testing scored 85/100. The companion security scan produced 87/100 with outcome PASSED. Overall trust score: 87/100. Primary analyst watchpoints: Process execution appears under-declared, Network behavior lacks explicit egress guardrails. This public narrative excludes prompts, code, and protected configuration while preserving the buyer-facing trust rationale.