Access, host changes, and teardown
Use the rules in SKILL.md and the live-target checks in core.md. Read only the matching deployed release README sections for the requested operation. These are targeted operations using the same discover, plan, approve, execute, and verify approach; do not restart deployment preparation.
Access
Use "After deployment", "Windows jumpbox (optional)", and "Other ways to allow RDP". Present only relevant access paths. Prefer an existing organizational managed prefix list, then a user-selected IP; offer the documented Session Manager path where applicable. Never open RDP to the world.
Discover resources and specify any security-group/IAM change, access source, and later rule removal in the agreed plan. Give resolved credential-retrieval instructions for the user to run themselves.
Host changes
Use "Adding or removing hosts", including the documented VCF-side prerequisites. Identify the environment and host, plan the exact change, perform the approved documented action, and verify the result.
Teardown
Use "Teardown", "Safer order", and "What's left after teardown". Identify environment, associated stacks, and removal scope. Follow the entry skill's staged approval/removal rule. Verify environment removal completed before proposing infrastructure removal; an empty result for an ambiguous target is insufficient evidence.
Before proposing a teardown, list what the release keeps by default, whether each still bills, and the option that removes it. Report retained resources, continuing charges, warnings, and the documented manual cleanup that remains; the check that nothing is left billing covers the kept resources too. Update the shared record with observed results.