Skills still instruct 'agent-relay fleet enable', which always fails since fleet became always-on

[khaliqgant]

Problem

Two skills tell agents to run agent-relay fleet enable before bringing nodes up, saying fleet is "off by default":

• skills/orchestrating-agent-relay/SKILL.md ~L682-686 ("Enable fleet nodes for the workspace FIRST — it is off by default…")
• skills/multi-host-live-mount/SKILL.md ~L614

On agent-relay 12.4.1 that command always fails (RelaycastMessagingClient.workspace.fleetNodes requires @relaycast/sdk with the workspace fleet nodes API.). relaycast#212 removed the per-workspace fleet flag, and fleet is now always on (tracked in AgentWorkforce/relay#1840). An agent following either skill hits the error on its first step. It can't tell whether the failure matters (it doesn't: nodes register and take spawns without it).

Verified live 2026-09-22: two nodes on a fresh workspace registered and accepted targeted spawns with no fleet enable.

Fix

• Remove the fleet enable step and the "off by default" claim from both skills. If anything is kept, say fleet nodes are always on.
• orchestrating-agent-relay: also drop or fix any fleet config / fleet disable / fleet inherit references (same dead API).
• Bump each changed skill's version in prpm.json (and the README table).

connecting-agents-across-machines (#109) already follows the corrected flow and can be the reference.

Acceptance

git grep -n "fleet enable\|fleet config\|fleet disable\|fleet inherit" skills/ returns nothing, or only a note that the commands are no longer needed.

[khaliqgant]

Also in scope: orchestrating-agent-relay troubleshooting (~L899) and its placement example (~L747) say targeted fleet spawn needs RELAY_AGENT_TOKEN. On 12.4.1 targeted spawns worked live with no RELAY_* env, and the requires an agent token error no longer exists in relay source. Drop the requirement, or make it optional.