fix: resolve lint warnings across server and web#4
Merged
Conversation
- Remove unused imports: ProviderApprovalDecision, ProviderSendTurnInput, CreateAgentSessionOptions, ProviderUserInputAnswers (PiAdapter.ts); ProviderInstanceId (serverRuntimeStartup.ts + test); HttpBody (server.test.ts) - Rename unused variables to underscore-prefixed: _piEnvironment (PiAdapter.ts), _eventLoggers (PiDriver.ts) - Remove unused encodeJsonString declaration (PiTextGeneration.ts) - Add composerRef to useCallback/useEffect deps in ChatView.tsx (exhaustive-deps: focusComposer, addTerminalContextToDraft, keyboard shortcut handler, onRespondToActivePendingUserInputChoice, onChangeActivePendingUserInputCustomAnswer, onSubmitPlanFollowUp, onImplementPlanInNewThread) - Replace spread-in-map with Object.assign in clientPersistenceStorage.ts (oxc no-map-spread) - Extract throwRegistryReadUninitialized to module scope in catalog.test.ts (unicorn consistent-function-scoping) - Opt into Node.js 24 for release job to silence action-gh-release deprecation
Track activeTextItemId/activeReasoningItemId on PiTurnState so text content is emitted as proper item.started/content.delta/item.completed sequences rather than bare content.delta events without an itemId. Boundaries are closed when a tool execution starts or the turn ends, giving the UI the structure it needs to display text interleaved with tool calls instead of merging all text together. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Remove hardcoded BUILT_IN_MODELS from PiProvider. Models are now discovered at runtime by running `pi --list-models` in the background enrichSnapshot hook, surfacing every provider/model pair Pi has access to (anthropic, cursor, openai, etc.) using `provider/model` slugs. - Add parsePiListModelsOutput() to parse the tabular CLI output - Add discoverPiModels effect (12s timeout) returning all rows - Add enrichPiSnapshot (mirrors enrichCursorSnapshot pattern): chains version-advisory enrichment then model discovery, publishes updated snapshot with discovered + custom models - checkPiProviderStatus and makePendingPiProvider now start with an empty built-in list; enrichment fills in the real list - Wire enrichPiSnapshot into PiDriver enrichSnapshot hook Cursor models (and any other provider models) now appear automatically after login without restarting t3code.
…napshot Skills discovered from the filesystem are now also exposed as /skill:<name> slash commands so they appear in the "/" menu alongside built-in commands. A /compact command is added with direct piSession.compact() dispatch, and $skill references from the "$" menu are translated to /skill: format for Pi's native expansion. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Version check and skill discovery now run in parallel via Effect.all and are cached with a 5-minute TTL using the same Cache.make pattern as ClaudeDriver, eliminating redundant subprocess spawns and filesystem scans on repeated snapshot refreshes. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The CLI writes its model table to stderr, not stdout. Fixes model discovery returning zero results.
The adapter was restructured to use RPC commands but the /compact interception and $skill→/skill: translation were not carried over. This re-adds both: /compact dispatches a compact RPC command directly, and $skillname references are rewritten to /skill:skillname for Pi's native skill expansion. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…heck Spawns a short-lived `pi --mode rpc` process during the capabilities probe and sends `get_commands` to discover extension commands, skill commands, and prompt templates that Pi knows about. These are surfaced as slash commands in the provider snapshot alongside the static /compact entry. Falls back gracefully if the probe times out or fails. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The RPC get_commands response already returns names like "skill:lsf", so wrapping with another "skill:" produced "skill:skill:lsf". Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…ations Skill commands sent via / without instructions cause Pi to hang because the LLM receives a skill block with no user query. Skills remain discoverable via the $ menu where users naturally add instructions. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The $ menu reads from the skills array which was only populated by filesystem scanning. RPC-discovered skill commands are now merged into the skills array so they appear in the $ menu where users can add instructions alongside the skill invocation. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The RPC returns names like "skill:checkin" but the frontend title-cases the name field, producing "Skill Checkin". Strip the prefix so skills display as just "Checkin", "Glab", etc. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
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
1 participant
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 Changed
Why
UI Changes
Checklist