Share AI work without making it publicly discoverable
Public-by-link pages make reviews and handoffs fast, but safe sharing still depends on clear boundaries and deliberate content choices.
Public by link is a sharing model
A public-by-link page removes account friction. Anyone with the exact URL can open the artifact, which is useful for reviews, handoffs, client previews, and async collaboration across tools.
The page is not placed in a public directory or user-page sitemap, and it requests no indexing. Those choices reduce casual discovery. They do not turn the URL into an access-control boundary.
Unlisted does not mean private
A link can be forwarded, logged, included in analytics, or pasted into another system. Treat every published page as content that may reach someone beyond the first recipient.
Never publish credentials, access tokens, private customer data, production secrets, or material whose exposure would create harm. Keep sensitive context in systems designed for confidential access control.
Publish the review layer
Share the plan, synthesis, decision, status, or artifact that someone needs to review. Avoid copying raw internal data simply because it helped the agent produce the result. The review layer should contain the minimum useful context for its audience.
This separation also improves clarity. A stakeholder usually needs the conclusion and its supporting rationale, not every private source the agent consulted.
Rotate when the audience changes
Stable links are valuable while the audience remains appropriate. If a URL reaches the wrong person or the sharing context changes, rotate the public link and distribute the replacement intentionally.
Fast sharing and careful boundaries can coexist. The key is understanding that discoverability controls reduce exposure while authorization controls access—and choosing the right system for each kind of content.
