v0.4.6

# v0.4.6 — HTML/PDF Previews, Event Adapters & OSS Bug Fixes

Major feature release adding inline HTML and PDF preview rendering, a new unified event adapter architecture for multi-backend support, batch session operations, and fixes for several community-reported issues.

---

## Features

- **HTML & PDF inline previews** — New `html-preview` and `pdf-preview` code blocks render rich HTML content and PDF documents directly in chat, with multi-item tab support and fullscreen overlay viewing (0d22069b)
- **Batch session context menu** — Right-clicking during multi-select now shows batch operations (status, labels, flag, archive, delete) instead of single-session actions (61759915)
- **Dynamic Codex model discovery** — Fetch available models from Codex app-server with 30-min periodic refresh, falling back to hardcoded registry on failure. Support for multiple connections of the same provider type with unique slug generation (1bcb1083)
- **Mustache template engine** — New template system for dynamic content rendering with `{{variable}}` syntax, partials, conditionals, and loops (a779bf83)
- **Source templates** — Sources can now ship HTML templates for branded, consistent rendering of their data via the new `render_template` tool. Templates use Mustache syntax with self-describing metadata headers (992a1a50)
- **Fast mode feature flag** — Moved to shared package for cross-package availability (e8295580)

## Bug Fixes

- **Chat history empty on macOS login launch** — Added Promise-based initialization gate to SessionManager; `GET_SESSIONS` handler now awaits init completion to prevent returning empty results during startup race. Fixes [#254](#254) (9ed91fbf)
- **Window not draggable across all panel headers** — App menu, session list panel, and panel headers now have consistent drag regions so the window can be moved from any header area. Fixes [#255](#255) (e1ed6619)
- **Skills can't auto-activate required sources** — Added `requiredSources` field to SKILL.md frontmatter; skills now declare source dependencies that are auto-enabled before the agent's first turn. Fixes [#249](#249) (9ed91fbf)
- **File links in chat not clickable** — csv, pdf, png, and other file types were not matched by the link click handler. Extension list is now derived from `file-classification.ts` to stay in sync automatically (a3fa35e4)
- **source_test reports expired for valid OAuth tokens** — Now attempts token refresh before reporting expired, matching the actual connection pipeline behavior (7a9bc9d9)
- **Interrupted responses cause LLM confusion** — Partial content from stopped/redirected responses now gets a context note injected so the model doesn't misinterpret follow-up messages (8fd1eb7d)
- **Oversize icons in workflow selectors and dropdowns** — Fixed icon sizing in StyledDropdown and workflow selector components (2c917b88, aef87854)
- **Windows Git Bash path not persisted** — Detected `CLAUDE_CODE_GIT_BASH_PATH` is now saved to config and restored on startup (9ed91fbf)
- **Per-session env override race condition** — Replaced global mutable `optionsEnv` with per-session `envOverrides` to fix `ANTHROPIC_BASE_URL` race across concurrent sessions (1bcb1083)

## Improvements

- **Unified event adapter architecture** — Extracted common streaming/event logic into `BaseEventAdapter` with backend-specific adapters for Claude, Codex, and Copilot. Includes new `EventQueue` for ordered event delivery (multiple commits)
- **TodoState → SessionStatus rename** — Consistent naming across 54+ files with backward compatibility for existing hook configs and session file migration (e5987124)
- **Block spacing improvements** — Better visual spacing between content blocks in chat (a0402076)
- **Transform tool timeout improvements** — More robust timeout handling for data transformation scripts (c03011b6)

Author: github-actions[bot]

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -0,0 +1,123 @@
+name: Bug Report
+description: Report a bug or unexpected behavior in Craft Agents
+labels: ["bug"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Thanks for taking the time to report a bug! Please fill out the sections below so we can reproduce and fix the issue.
+
+  - type: input
+    id: version
+    attributes:
+      label: Craft Agents Version
+      description: "Found in Settings or the title bar (e.g. 0.4.6)"
+      placeholder: "0.4.6"
+    validations:
+      required: true
+
+  - type: dropdown
+    id: os
+    attributes:
+      label: Operating System
+      options:
+        - macOS (Apple Silicon)
+        - macOS (Intel)
+        - Windows 11
+        - Windows 10
+        - Linux (Ubuntu/Debian)
+        - Linux (Fedora/RHEL)
+        - Linux (Arch)
+        - Linux (Other)
+    validations:
+      required: true
+
+  - type: input
+    id: os_version
+    attributes:
+      label: OS Version
+      description: "e.g. macOS 15.3, Windows 11 24H2, Ubuntu 24.04"
+      placeholder: "macOS 15.3"
+    validations:
+      required: true
+
+  - type: dropdown
+    id: ai_provider
+    attributes:
+      label: AI Provider
+      description: Which AI provider/connection are you using?
+      options:
+        - Anthropic API (direct)
+        - Anthropic API (custom endpoint)
+        - OpenAI / Codex
+        - Copilot (GitHub)
+        - Other
+    validations:
+      required: true
+
+  - type: input
+    id: model
+    attributes:
+      label: Model
+      description: "Which model are you using? (e.g. Claude Opus 4.6, GPT-4.1)"
+      placeholder: "Claude Opus 4.6"
+
+  - type: textarea
+    id: description
+    attributes:
+      label: Description
+      description: A clear description of what the bug is.
+    validations:
+      required: true
+
+  - type: textarea
+    id: steps
+    attributes:
+      label: Steps to Reproduce
+      description: Step-by-step instructions to reproduce the behavior.
+      value: |
+        1.
+        2.
+        3.
+    validations:
+      required: true
+
+  - type: textarea
+    id: expected
+    attributes:
+      label: Expected Behavior
+      description: What did you expect to happen?
+    validations:
+      required: true
+
+  - type: textarea
+    id: actual
+    attributes:
+      label: Actual Behavior
+      description: What actually happened?
+    validations:
+      required: true
+
+  - type: textarea
+    id: screenshots
+    attributes:
+      label: Screenshots / Screen Recordings
+      description: |
+        For **UI issues**, please attach screenshots or screen recordings showing the problem.
+        You can drag and drop images/videos directly into this field.
+
+  - type: textarea
+    id: logs
+    attributes:
+      label: Debug Logs
+      description: |
+        For **non-UI issues** (crashes, errors, connection problems), please attach relevant logs from a debug session.
+
+        Launch the app with `-- --debug` and reproduce the issue. See the [Troubleshooting section in the README](https://github.com/lukilabs/craft-agents-oss#troubleshooting) for platform-specific commands and log file locations.
+      render: shell
+
+  - type: textarea
+    id: additional
+    attributes:
+      label: Additional Context
+      description: Any other context about the problem (MCP sources used, workspace config, etc.)

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -0,0 +1,36 @@
+name: Feature Request
+description: Suggest a new feature or enhancement
+labels: ["enhancement"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Have an idea for Craft Agents? We'd love to hear it!
+
+  - type: textarea
+    id: problem
+    attributes:
+      label: Problem or Motivation
+      description: What problem does this feature solve, or what workflow does it improve?
+    validations:
+      required: true
+
+  - type: textarea
+    id: solution
+    attributes:
+      label: Proposed Solution
+      description: Describe the feature or change you'd like to see.
+    validations:
+      required: true
+
+  - type: textarea
+    id: alternatives
+    attributes:
+      label: Alternatives Considered
+      description: Any workarounds or alternative approaches you've tried.
+
+  - type: textarea
+    id: additional
+    attributes:
+      label: Additional Context
+      description: Screenshots, mockups, links, or any other context.

README.md

@@ -342,7 +342,7 @@ Or configure manually in `~/.craft-agent/workspaces/{id}/hooks.json`:
 - **Command hooks** — run shell commands with event data as environment variables (`$CRAFT_LABEL`, `$CRAFT_SESSION_ID`, etc.)
 - **Prompt hooks** — create a new agent session with a prompt (supports `@mentions` for sources and skills)
 
-**Supported events:** `LabelAdd`, `LabelRemove`, `PermissionModeChange`, `FlagChange`, `TodoStateChange`, `SchedulerTick`, `PreToolUse`, `PostToolUse`, `SessionStart`, `SessionEnd`, and more.
+**Supported events:** `LabelAdd`, `LabelRemove`, `PermissionModeChange`, `FlagChange`, `SessionStatusChange`, `SchedulerTick`, `PreToolUse`, `PostToolUse`, `SessionStart`, `SessionEnd`, and more.
 
 See the [Hooks documentation](https://agents.craft.do/docs/hooks/overview) for the full reference.
 

apps/electron/electron-builder.yml

@@ -24,6 +24,7 @@ files:
   # Include network interceptor for API error capture and MCP schema injection
   - packages/shared/src/network-interceptor.ts
   - packages/shared/src/interceptor-common.ts
+  - packages/shared/src/feature-flags.ts
   # Include bundled Bun runtime (platform-specific, set by build script)
   - vendor/bun/**/*
   # Include bundled Codex binary (platform-specific, downloaded by build script)

apps/electron/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@craft-agent/electron",
-  "version": "0.4.5",
+  "version": "0.4.6",
   "description": "Electron desktop app for Craft Agents",
   "main": "dist/main.cjs",
   "private": true,

apps/electron/resources/docs/hooks.md

@@ -7,7 +7,7 @@ This guide explains how to configure hooks in Craft Agent to automate workflows
 Hooks allow you to trigger actions automatically when specific events occur in Craft Agent. You can:
 - Run shell commands when labels are added/removed
 - Execute prompts on a schedule using cron expressions
-- Automate workflows based on permission mode changes, flags, or todo state changes
+- Automate workflows based on permission mode changes, flags, or session status changes
 
 ## hooks.json Location
 
@@ -46,9 +46,11 @@ Hooks are configured in `hooks.json` at the root of your workspace:
 | `LabelConfigChange` | Label configuration changed | Always matches |
 | `PermissionModeChange` | Permission mode changed | New mode name |
 | `FlagChange` | Session flagged/unflagged | `true` or `false` |
-| `TodoStateChange` | Todo status changed | New status (e.g., `done`, `in_progress`) |
+| `SessionStatusChange` | Session status changed | New status (e.g., `done`, `in_progress`) |
 | `SchedulerTick` | Runs every minute | Uses cron matching |
 
+> **Note:** `TodoStateChange` is a deprecated alias for `SessionStatusChange`. Existing configs using the old name will continue to work but will show a deprecation warning during validation.
+
 ### Agent Events (passed to Claude SDK)
 
 | Event | Trigger | Match Value |