Human-Agent Protocol FAQs

Question 1

How does HAP ensure human oversight for AI agents?

Accepted Answer

HAP routes AI agent tool calls through its ticket system, requiring human approval or rejection via CLI or a desktop app. This provides full human oversight for AI actions, complemented by auditable event logs and configurable timeout behaviors.

Question 2

Is HAP's decision-making process auditable?

Accepted Answer

Yes, HAP maintains a comprehensive, append-only event log for all operations. This log utilizes SHA-256 hash chaining to ensure integrity verification, making every decision, action, and state change fully auditable and tamper-proof.

Question 3

What is Human-Agent Protocol (HAP)?

Accepted Answer

HAP (Human-Agent Protocol) is a universal protocol designed for symbiotic collaboration between humans and autonomous AI agents. It facilitates structured, auditable, and latency-safe management of AI agent decision requests through a ticket-based system.

Question 4

What problem does HAP solve for AI agents?

Accepted Answer

HAP prevents AI agents from blocking while waiting for human input. Agents create tickets with defined timeout behaviors (e.g., auto-approve, auto-reject, cancel), allowing them to continue operations or escalate gracefully without interruption, ensuring latency safety.

Question 5

Which AI platforms does Human-Agent Protocol integrate with?

Accepted Answer

HAP integrates seamlessly with major AI platforms and tools, including Claude, Codex CLI, OpenCode, VS Code Copilot, and Copilot CLI. It allows these platforms to leverage HAP's ticket-based approval system for tool execution.

Human-Agent Protocol

Human-Agent Protocol

소개

주요 기능

사용 사례

소개

주요 기능

사용 사례