← All use-case collectionsDecision collection

Build Your Own Agent Platform

Frameworks and extensible platforms used to assemble custom agents, tools, sandboxes, and orchestration.

Operator decisionWhich open-source agent foundation gives builders the clearest security and evidence controls?
Primary security questionWhich platform exposes usable isolation, permission, extension-trust, and evidence controls?
Evidence gate0 of 2 reviewed results available
Comparison held

No “best” claim yet.

This collection needs at least 2 eligible repositories with reviewed canonical results. Candidates remain visible for planning, but they are not ranked and no missing score is estimated.

CandidateEvidenceFunctional fitRequired authorityData sensitivityOperator burden
DeepSeek HarnessRepository ↗No reviewed resultcoreDeveloper-preview harness with high attention; exact commit and defaults still require static review.Plugin runtime, tools, permissions, and deployment infrastructurePrompts, workspace data, credentials, and infrastructure datahigh
DeerFlowRepository ↗No reviewed resultcoreExtensible platform with explicit orchestration and sandbox choices.Sub-agents, browser, sandbox, and research toolsWorkspace, retrieved content, and generated artifactshigh
OpenHandsRepository ↗No reviewed resultcoreMature platform with auditable isolation and permission decisions.Sandbox, shell, browser, repository, and integrationsSource code, task data, and credentialshigh

Campaign rationale

Foundation choices create durable security consequences before application controls exist.

What this page does not claim

Membership, stars, forks, discussion volume, and functional fit do not change a repository’s static score or assurance level.

Workflow boundary

A complete stack earns no workflow label until its exact components, permissions, prompts, tools, environment, and scenarios have been assessed together.