Should captured information become agent memory automatically?

Disclosure: I’m building ChaseOS Studio, and this is the memory boundary I’m testing.

Most agent-memory stacks make capture and trust almost the same event: ingest a page or attachment, embed it, and let future agents retrieve it.

I’m taking a different approach. New material enters Intake first. Its source remains visible, a person decides whether it belongs in the workspace graph, and rejection happens before it becomes durable agent context.

That separates three decisions that are often collapsed:

• Is this useful enough to capture?

• Is it trustworthy enough to remember?

• Does the agent have authority to act on it?

The trade-off is extra friction. The benefit is that an untrusted page, stale document or generated inference cannot quietly become durable memory and influence later work.

For people building agent products: where would you place this boundary? Are there sources you would admit automatically, and what evidence would you require?

If you want to inspect the current Intake and approval model, ChaseOS Community is the complete free local Windows product:

https://chaseos.ai/?utm_source=producthunt&utm_medium=forum&utm_campaign=governed_intake