Releases: devoxx/DevoxxGenieIDEAPlugin
Releases · devoxx/DevoxxGenieIDEAPlugin
v1.5.0
[1.5.0] - 2026-05-18
Added
- feat: Introduce langchain4j Skills and rename Custom Prompts → Commands (#1040)
- New
Skillssettings tab listing skills detected under~/.devoxxgenie/skills/,<project>/.devoxxgenie/skills/,.claude/skills/and.agents/skills/, with enable/disable, source, open-folder and reload actions SkillRegistryproject service loadsSKILL.mddefinitions; project skills override user skills (with warning)- Skills are wired into the agent tool chain via
AgentToolProviderFactoryand the system prompt fragment is appended byChatMemoryManagerwhen agent mode is on - Legacy
customPromptsXML state is migrated tocommandson load (settings tab renamed fromCustom PromptstoCommands)
- New
- feat(welcome): show active Skills on the welcome page above Quick Commands (#1041)
- feat(settings): replace six per-source
Open folderbuttons with a single button + source dropdown (#1041) - feat(skills): scan
.claude/skillsand.agents/skillsalongside.devoxxgenie/skills(#1041)
Fixed
- fix: fire
onCustomPromptsChangedfromSkillsSettingsConfigurableon apply so the welcome page refreshes immediately (#1040)
Docs
- Add
Commands vs Skillsblog post and screenshot (#1040) - Update Commands rename and add Skills documentation (#1040)
- Fix duplicate footnote markers in
overview.md(#1040)
CI
- ci(claude-review): skip job on fork PRs (OIDC unavailable from forks)
- ci(lychee): exclude
intel.comfrom link check (403 on automated crawlers)
Contributors
v1.4.5
[1.4.5] - 2026-05-18
Added
- feat: Add copy button to user prompt bubble (#1024) (#1037)
- feat: add shell env file and shell selector for run_command tool (#1036)
- Add task-211: persistent semantic conversation memory backed by ChromaDB (#1028)
Fixed
- fix: surface multi-line run_command output in agent log preview (#1038)
- fix: support CJK characters in chat input field (#1034) (#1035)
- fix: destroy process tree in AcpTransport.close() to prevent thread leak (#1031)
- fix: revert Compose/Kotlin ecosystem bump that broke the UI (#1030)
Dependencies
Contributors
v1.4.4
[1.4.4] - 2026-05-14
Added
- Persistent semantic conversation memory backed by ChromaDB (task-211) (#1028)
Fixed
- Revert Compose/Kotlin/Skiko ecosystem bump to known-good versions and align JVM target to 21 (#1030)
- Restore last selected model state on restart (#1021) (#1022)
- Correct Devstral 2 model name (#1019)
- Apply IDE Editor font to prompt input field (#1013)
- Prevent ACP reader thread leak on transport close
Dependencies
- Bump the gradle-dependencies group with 38 updates (#1023)
- Bump @babel/plugin-transform-modules-systemjs in /docusaurus (#1026)
- Bump fast-uri from 3.0.6 to 3.1.2 in /docusaurus (#1025)
- Bump follow-redirects from 1.15.11 to 1.16.0 in /docusaurus (#1011)
Contributors
- @app/dependabot
- @flobo3
- @mihaibuba
- @stephanj
V1.4.3
[1.4.3]
Added
- feat(welcome): blog feed on welcome screen + friendlier analytics copy (#1010)
- feat(analytics): track feature enablement & usage (task-209) (#1008)
- feat(analytics): anonymous LLM provider/model usage analytics (#1005)
- feat: add Exo distributed AI cluster as LLM provider (#1002)
Fixed
- fix(analytics): harden offline tracking delivery (#1007)
Dependencies
Contributors
- @app/dependabot
- @stephanj
[1.4.2]
Added
- feat(analytics): track feature enablement & usage (task-209) (#1008)
- feat(analytics): anonymous LLM provider/model usage analytics (#1005)
- feat: add Exo distributed AI cluster as LLM provider (#1002)
Fixed
- fix(analytics): harden offline tracking delivery (#1007)
Dependencies
Contributors
- @app/dependabot
- @stephanj
v1.4.1
What's Changed
Updated
- Upgrade 40 Gradle dependencies including Langchain4J to 1.12.2
- Update LLM models.json with latest provider model names (Anthropic 1M context for 4.6 models, new Grok 4.20 lineup, Groq gpt-oss models, Bedrock Claude 4.6/4.5/4.1 and Llama 4)
- Add ModelConfigJsonTest to validate models.json schema and parsing
Fixed
- Rollback Kotlin/Compose/Skiko versions to fix Compose UI rendering crash (UnsatisfiedLinkError in Metal renderer, NoSuchMethodError in Kotlin Duration stdlib)
- Fix SafeComposeContainer to call super.addNotify() before adding children so native init errors are caught properly
- Fix Kotlin compilation errors for multiplatform-markdown-renderer API changes
- Fix AnthropicChatModelTest for removed model name constant in Langchain4J 1.12.2
Full Changelog: v1.4.0...v1.4.1
v1.4.0
[1.4.0]
Added
- Exo distributed AI cluster as new local LLM provider — run large AI models across multiple Apple Silicon devices connected via Thunderbolt (#1002)
- Exo auto-discovery of downloaded models from /state API
- Automatic Exo instance creation with placement preview across cluster
- Background progress bar during Exo model loading with cancellation support
- Auto-recovery when Exo instances disconnect or get recycled
- Collapsible cluster status panel above chat showing nodes, memory, GPU usage, temperature, and active instance status
- Exo Docusaurus documentation page

v1.3.3
v1.3.2
[1.3.2]
Added
- Integrate Spec-to-Code panels into main DevoxxGenie tool window instead of separate sidebar (#987)
- Force Skiko software rendering on Windows via AppLifecycleListener (#986)
- Include archived tasks in backlog duplicate detection (#985)
Fixed
- File attachment double-click not adding file to window context (#988)
- Remove flaky OpenRouterChatModelFactory tests
- Defensive copy in ProcessedCompletion record constructor
- Refactor Kotlin Compose compiler arg filter to use compilerOptions API
Changed
- Centralize system prompt construction in ChatMemoryManager
Documentation
- Update FAQ and troubleshooting for automatic Windows software rendering
v1.3.1
[1.3.1]
Added
- Bedrock bearer token authentication support (#983)
- 'Force software rendering' setting for Skiko/GPU issues (#981)
Fixed
- Separate Ollama context window override from discovered metadata (#979) —
num_ctxis now only sent when explicitly enabled via new "Ollama Request Context Override" setting - Catch Skiko/Direct3D init errors to prevent chat UI crash (#981)
Documentation
- Add troubleshooting guide and FAQ entry for GPU rendering issues (#981)
- Add confirmation popup screenshot to Event Automations docs page
v1.3.0
[1.3.0]
Added
- Event Automations (BETA) — AI agents triggered by IDE events (file open/save, build failure, test failure, commit, process crash)
- Built-in agents: Code Review, Build Fix, Debug, Test Generator, Code Explainer, plus Custom agents
- Event Automation settings panel with enable/disable, custom prompts, and auto-run per mapping
- Template engine for agent prompts with
{{context}},{{content}},{{files}},{{event}},{{timestamp}},{{meta.KEY}}variables - Before-commit automation includes full staged diff (before/after file content) for line-level code review
- "More Info" help link in Event Automations settings panel linking to documentation
- Event Automations Docusaurus documentation page with settings screenshot
Fixed
- Event automations ensure DevoxxGenie tool window is initialized before dispatching prompts (prevents silent prompt loss)
TestExecutionListeneruses project-awaretestSuiteFinished(root, project)overload for correct project in multi-project sessions- Removed 10 unsupported
IdeEventTypeenum values and 3AgentTypeentries that had no backing listeners - CLI runner error message readability on test connection failure (#971)
Testing
IdeEventTypeTestvalidates enum values match wired listeners to catch future driftEventAutomationSettingsTestvalidates default mappings, agent/event type integrity, and boundary checks- Compose theme state retention test coverage