feat(agents): automatic model fallback when Ollama Cloud is exhausted

Keep ollama-cloud/* as primary. Each agent may declare fallback in
agents.json (xai-oc/grok-*). On quota/auth/provider failure, run-agent
switches once to the fallback and continues retries.

- agents.json: fallback fields for pm/junior/senior/qa/intern
- route.sh: emit fallback to GITHUB_OUTPUT
- agent.yml: pass FALLBACK env into run-agent
- install-opencode.sh: register primary+fallback models in provider maps
- run-agent.sh: failover on quota/credit/429/capacity-class errors

.gitea/workflows/agent.yml

@@ -222,6 +222,7 @@ jobs:
           SELF_TOKEN: ${{ steps.prep.outputs.name == 'pm' && secrets.TOKEN_PM || steps.prep.outputs.name == 'junior' && secrets.TOKEN_JUNIOR || steps.prep.outputs.name == 'senior' && secrets.TOKEN_SENIOR || steps.prep.outputs.name == 'lead' && secrets.TOKEN_LEAD || steps.prep.outputs.name == 'qa' && secrets.TOKEN_QA || steps.prep.outputs.name == 'ops' && secrets.TOKEN_OPS || steps.prep.outputs.name == 'intern' && secrets.TOKEN_INTERN || '' }}
           NAME: ${{ steps.prep.outputs.name }}
           MODEL: ${{ steps.prep.outputs.model }}
+          FALLBACK: ${{ steps.prep.outputs.fallback }}
           VISION: ${{ steps.prep.outputs.vision }}
           MODE: ${{ steps.prep.outputs.mode }}
           WORKMODE: ${{ steps.prep.outputs.workmode }}   # build | discuss (devs consulted in-thread)