Repository navigation
Task thinking effort, task rename, English dark-mode asset graph, New token usage default - #1
Merged
Merged
Conversation
createTask accepts thinking_effort=auto, binding the new task to the active model's provider-default variant (no thinking fields sent) so easy work doesn't pay for max reasoning; unset/max leaves the task on the max-thinking rotation pool. Bridge exposes the same field plus rename_task, and the create_task description now separates display name from long context.
…lts to New The graph legend/labels were Chinese and the canvas stayed light inside dark mode. Graph colors now derive from the app's CSS theme tokens and redraw on theme change, preserving layout and zoom. The LLM Token Usage panel opens on New, with Legacy still selectable.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Four user-reported needs, one branch:
thinking_effort: "auto"on task creation binds the task to the active model's provider-default variant (no thinking fields sent), while unset/maxleaves the task on the max-thinking rotation pool. Exposed in the iMessage bridgecreate_tasktool so easy jobs get cheaper reasoning and hard jobs keep max. Deliberate: task-levelmaxnever injects an effort a given model may reject — the profile's configured ceiling always wins.rename_task(existingPATCH /api/tasks/{id}), andcreate_tasknow spells out thatnameis the display title anddescriptionis long-form context — names were landing in descriptions. Task names are optional everywhere; seeding reads Name, then Goal, never Description; creation name is snapshotted as durable scope so later renames never change what a task targets.Validation
go build ./...;go test ./server -run 'TestPickAutoProfile|TestHeadRank|TestTaskRename',go test ./llmpool ./agent— pass (rename test also run against real Postgres on the server box).npm run build:static;node --test src/lib/graph-theme.test.mjs— pass.no-mistakesreview applied 20+ gate fixes across 20 rounds (applied-config consistency, seeding rules, limiter sharing, schema migrations). The final review round is parked: codex hit its usage limit, claude/pi lack valid credentials. Re-runno-mistakes axi runon this branch once credits reset to finish the gate.Not in scope (recorded)