fix(gameplay): use universe-generated AI companions
Some checks failed
CI / validate (pull_request) Waiting to run
CI / validate (push) Failing after 34m2s

Co-authored-by: multica-agent <github@multica.ai>
This commit is contained in:
2026-08-22 14:50:14 +05:00
parent 56bddafbe8
commit f6c0fbcfeb
12 changed files with 632 additions and 124 deletions

View File

@@ -58,7 +58,7 @@ If the worker reports `POST /rest/v1/rpc/claim_ai_job 404` and `/rest/v1/profile
1. Click **Enter the Alpha** to open registration. Every account requires a display name, email, password, and 13+ confirmation. Existing users sign in with email/password; forgotten passwords use the email recovery flow.
2. Create a world in the dynamic coauthor chat. Unfinished conversations and generated drafts appear on the dashboard and resume after a reload.
3. Review every starting-world field, including the owner-only hidden threat, then confirm it.
4. A new campaign starts immediately with an owner-controlled hero and a persistent AI companion. Additional players create a human character manually or ask the coauthor for an editable draft; owners can add more AI companions the same way.
4. During world creation, the coauthor also creates three editable, universe-specific AI companions. A new campaign starts immediately with those companions and an owner-controlled hero. Additional players create a human character manually or ask the coauthor for an editable draft; owners can add more AI companions the same way.
5. The owner creates an expiring private invite link. Signed-in email users join through `/join/:token`.
6. Players save actions and mark them ready. The final ready action queues the round automatically; the owner can also continue without waiting.
7. The worker resolves server-owned rolls and state, publishes narration, and opens the next round. Visible campaign tabs synchronize every two seconds and immediately when the tab regains focus.