The safeguards map / 3.2 Agent harnesses

Agent Development Kit

Google (company). No foundation, released. Apache-2.0. Mapped; as of .

Drafted with AI

Drafted with AI from the cited sources under the direction of UnGovr staff. UnGovr holds editorial responsibility for this page. How this site is made

An agent framework whose callbacks can skip a model or agent step and whose tool confirmation pauses a tool for a yes or no from a person.

Where it sits

3.2 Agent harnesses (primary): Runs developer-defined callbacks at set points in the agent's run that can replace or skip the next step, and pauses a tool for confirmation from a person or supervising system.

They allow you to observe, customize, and even control the agent's behavior at specific, predefined points without modifying the core ADK framework code.

captured

Useful for handling simple requests directly or enforcing access control.

captured

Some agent workflows require confirmation for decision making, verification, security, or general oversight.

captured

Facts

Concern
safetysecurity
Names the law
no
Subject
agent
Used by
those who make it
Runs
at runtime
Policy comes from
owner-entered settings
Can the agent bypass it
depends

Handoffs

produces Runtime records: OpenTelemetry traces (OTLP exporter, documented for Kotlin)

Every quote is verbatim from a stored capture of the page linked under it. Placement is our reading of the product's own documentation; no publisher has reviewed or endorsed it. Rows CC BY 4.0. Names, marks and logos belong to their owners; logos identify each product and are not part of the dataset.