When sharing screenshots from a real cloud-server AI/Agent environment, treat screenshots as sensitive operational artifacts.
- app IDs, app secrets, tokens, refresh tokens
- user IDs, open IDs, tenant IDs, union IDs
- folder tokens, document tokens, task IDs
- absolute local paths when they reveal private environment structure
- remote URLs that expose sensitive object IDs
- message IDs, log IDs, troubleshooting URLs containing request identifiers
- terminal history that includes private commands or local-only file names
- server hostnames and public IPs
- private project names
- customer names, employee names, internal group names
- chat IDs, wiki space IDs, folder names tied to real organizations
- browser tabs or bookmarks visible in the screenshot
- Capture the screenshot.
- Review it at full size before publishing.
- Redact identifiers and sensitive paths.
- If the screenshot includes a command line, check for secrets in wrapped lines.
- If the screenshot includes a notification or sync summary, redact runtime metadata such as doc IDs and folder tokens.
- Prefer screenshots from demo directories and demo remote roots.
- Prefer
--dry-runoutput where possible. - Prefer docs written from example configs rather than real environment output.
If a screenshot would help someone locate your server, identify your tenant, or access a specific remote object, redact it before publishing.