Show HN: Joy – AI 에이전트를 위한 오픈 신뢰 네트워크 (AI-대-AI 보증)

Source: Dev.to

Overview

Show HN: Joy — open trust network for AI agents (AI-to-AI vouching). Agents can vouch for other agents to delegate actions. Read the project: (HN thread: )

Idea in one line

Let agents carry machine‑signed claims so downstream agents accept actions without human friction.

Benefits and Risks

• Benefit: Faster pipelines.
• Risk: Bigger blast radius—one compromised agent can trigger PRs, SQL queries, or payments.

Practical Controls

You can ship today:

• BYOK (your API keys)
• Mandatory human approval for publish/transactions
• Per‑action scopes
• Short‑lived signed claims
• Revokable attestations
• Append‑only audit logs

I run an event‑driven Claude → GitHub pipeline; these cut risk.

Test to Run

In 10 minutes:

1. Spin a vouching agent.
2. Revoke its key.
3. Observe if downstream still accepts its vouches.

Measure detection time and impact. If you serve legal/finance clients, treat human review as a product feature, not an optional checkbox.

Call for Feedback

Thoughts from builders?