ENG-648: Reduce token usage — prompt caching, prompts→skills, cheap router tier#239

Draft

torrmal wants to merge 13 commits into

jorge/eng-648-reduce-anton-cowork-server-token-usage-promptskills

Commits on Jul 8, 2026

ENG-648: token-efficiency analysis and plan (prompt caching, prompts→skills, cheap router tier)

torrmal
and
claude
committed
ENG-648 Phase 3: cheap front-model router (respond-or-delegate)

torrmal
and
claude
committed
ENG-648 Phase 2: move prompt tutorials into built-in skills

torrmal
and
claude
committed
ENG-648: rename the front-model gate to 'thalamus'

torrmal
and
claude
committed
ENG-660: route history summarization through the thalamus role

torrmal
and
claude
committed
ENG-660: name the config surface router_* (mechanism stays thalamus)

torrmal
and
claude
committed
ENG-660: normalize router_provider minds-cloud->openai-compatible

torrmal
and
claude
committed

Commits on Jul 10, 2026

Commits on Jul 16, 2026

Merge branch 'staging' into jorge/eng-648-reduce-anton-cowork-server-token-usage-promptskills
ea-rus
committed

Commits on Jul 17, 2026

Merge branch 'staging' into jorge/eng-648-reduce-anton-cowork-server-token-usage-promptskills
ea-rus
committed

Commits on Jul 20, 2026