Seat a Token Holder Before Shared Model Runs

Description and use

Seat a token holder for shared model servers to ensure a single active session, preserve context, and enforce traceability of work.

Shared free coding servers lose session history and context when no named owner is assigned, leading to chaotic, untraceable model runs.

Ensures only one live session per shared lane, preserves context, and enforces traceability of model runs.

1. Run scripts/seat-run-token.sh "goal text" to create a run token and receipt file.
2. Paste the printed RUN_TOKEN into the wiki line Live stamp.
3. Close any leftover chat or server session before handing off.
4. Incoming holder pastes stamp path and last command in the ticket.
5. Incoming holder reruns scripts/seat-run-token.sh --resume <id> to resume.
6. Incoming holder writes seated on the wiki line.
7. CI check scripts/require_run_token.py verifies RUN_TOKEN= in PR body before merge.

The repair is a run token, not a clever prompt. Only the token holder may start a model session.

The merge steward can reject unmarked work in CI. A tiny check looks for RUN_TOKEN= in the pull request body.