openagentd 1.76.2__tar.gz → 1.78.0__tar.gz

openagentd-1.78.0/.openagentd/skills/oad/debug/SKILL.md

@@ -0,0 +1,56 @@
+---
+name: oad/debug
+description: OpenAgentd workflow for investigating bugs, regressions, sessions, and runtime issues.
+---
+
+Debug the reported issue across any surface of the OpenAgentd stack.
+
+## 1. Triage the report
+
+Extract: symptom, expected behavior, reproduction steps, **affected surface** (backend / frontend / desktop / mobile / agent / provider), session id, workspace, model, logs, and timing clues.
+
+If the report is ambiguous, inspect available evidence first; ask only when a missing decision blocks safe progress.
+
+---
+
+## 2. Route to the right reference
+
+Load the surface-specific reference for deeper commands, file maps, and gotchas:
+
+- **Backend / API / agent / provider** → `oad/debug/reference/backend`
+- **Frontend (web UI)** → `oad/debug/reference/frontend`
+- **Desktop or mobile (Tauri / Rust)** → `oad/debug/reference/tauri`
+
+When the issue spans multiple surfaces, load all relevant references.
+
+---
+
+## 3. Reproduce narrowly
+
+- Recreate the smallest scenario that demonstrates the bug.
+- Match the user's mode / workspace / model / message sequence when relevant.
+- Capture durable evidence: raw HTTP response, persisted history, SSE events, logs, failing test output, or UI state snapshot.
+
+---
+
+## 4. Diagnose from code and evidence
+
+- Search for existing patterns before editing.
+- Identify the boundary that failed: route validation, persistence, queueing, stream emission, agent loop, hook, tool, provider, frontend store, renderer, or Rust process.
+- Preserve unrelated work; do not reset or overwrite changes you did not make.
+
+---
+
+## 5. Fix surgically
+
+- Make the smallest change that addresses the proven root cause.
+- Add or update focused regression coverage at the closest layer.
+- Update related docs only when behavior, API contract, or operator workflow changed.
+
+---
+
+## 6. Verify and report
+
+- Re-run the reproduction and focused tests / checks for the touched areas.
+- If feasible, run the repository's standard lint / type / test commands for the changed surface.
+- Report: root cause, changed files, checks run with results, and any remaining risk or unverified area.

openagentd-1.78.0/.openagentd/skills/oad/debug/reference/backend.md

@@ -0,0 +1,66 @@
+# Debug reference: Backend / API / agent / provider
+
+Use when the symptom is in API routes, persistence, queueing, SSE streaming, agent loops, tool execution, or provider calls.
+
+---
+
+## Evidence commands
+
+Start with no-server diagnostics, then live ones:
+
+```bash
+uv run python -m manual.backend_log              # scan app.log for WARNING/ERROR (no server needed)
+uv run python -m manual.health                   # server liveness + agent roster
+uv run python -m manual.team_history <SESSION_ID>
+uv run python -m manual.team_timeline <SESSION_ID> --full
+uv run python -m manual.team_sse "message" --session <SESSION_ID>
+```
+
+See `manual/AGENTS.md` for the full catalogue — every script supports `-h`.
+
+Provider smoke tests:
+
+```bash
+uv run python -m manual.try_providers.<provider>  # e.g. manual.try_providers.anthropic
+```
+
+---
+
+## File map
+
+```
+app/
+  server.py            FastAPI app entry, route registration
+  api/                 Route handlers (sessions, agents, messages, tools, …)
+  core/                Business logic, agent runner, session manager
+  core/desktop_auth.py Desktop token validation
+  providers/           LLM provider adapters
+  db/                  SQLModel models, migrations, session factories
+manual/                Diagnostic scripts (no prod impact)
+tests/                 pytest coverage — match test to the layer you changed
+```
+
+---
+
+## Common failure boundaries
+
+| Boundary | What to inspect |
+|---|---|
+| Route validation | Pydantic schema, FastAPI dependency, HTTP status returned |
+| Persistence | SQLModel model, Alembic migration, `db/` session factory |
+| Queueing / ordering | Message queue in `core/`, SSE event emission order |
+| Agent loop | `core/` agent runner, tool dispatch, compaction logic |
+| SSE stream | `api/` stream route, `Emitter`, client reconnect behavior |
+| Provider call | `providers/` adapter, env vars, retry/timeout config |
+| Desktop auth | `core/desktop_auth.py`, token header, sidecar handshake |
+
+---
+
+## Verification
+
+```bash
+uv run pytest tests/ -x -q                      # full suite, stop on first fail
+uv run pytest tests/path/to/test.py -x -v       # focused
+uv run ruff check app/ && uv run ruff format --check app/
+uv run mypy app/
+```

openagentd-1.78.0/.openagentd/skills/oad/debug/reference/frontend.md

@@ -0,0 +1,83 @@
+# Debug reference: Frontend (web UI)
+
+Use when the symptom is in the React app — rendering, state, hooks, API calls, or live UI behavior.
+
+---
+
+## Evidence commands
+
+```bash
+cd web
+
+bun run typecheck          # TypeScript errors across all source
+bun run lint               # ESLint (catches unused vars, hook rules, a11y)
+bun test --reporter=verbose  # full test suite
+bun test <path/to/test>    # focused test file
+```
+
+For live UI issues (visual regression, interaction bugs, DOM/CSS state) load and follow the **`oad/frontend-debug`** skill — it drives the real UI via browser-use and can screenshot, inspect state, and simulate mobile keyboard.
+
+---
+
+## File map
+
+```
+web/src/
+  components/            UI components
+    ui/                  Zero-dep primitives (button, dialog, sheet, popover,
+                           dropdown, tooltip, switch, tabs — no shadcn/Base UI/CVA)
+      _use-deferred-unmount.ts  Exit-animation hook; must be called before any return null
+    AgentView/
+      UserBubble.tsx     User message bubble (mention highlighting, collapse, copy)
+    InboxBubble.tsx      Inter-agent inbox messages
+    InputBar.tsx         Composer (file attach, mentions, submit)
+    FloatingInputBar.tsx Floating variant
+    ToolCall/            Tool call display components
+    …
+  hooks/                 React hooks (session, streaming, settings, …)
+  stores/                Zustand stores (session, UI, pending messages, …)
+  queries/               TanStack Query factories (sessions, agents, messages)
+  api/                   Typed API client + generated types
+  utils/
+    LazyMarkdownBlock.tsx  Lazy markdown renderer (used in assistant + inbox bubbles)
+  routes/                TanStack Router pages
+  __tests__/             Vitest + RTL tests — mirror the component path
+```
+
+---
+
+## Common failure boundaries
+
+| Boundary | What to inspect |
+|---|---|
+| Render bug | Component file + its `__tests__/` counterpart |
+| State desync | Zustand store (`stores/`) + TanStack Query key factory |
+| API shape mismatch | `api/types.ts`, query/mutation in `queries/` |
+| Hook misfire | Custom hook in `hooks/` + React rules (StrictMode double-invoke) |
+| CSS / layout | Tailwind classes, `index.css` design tokens, dark-mode variants |
+| UI primitive bug | `components/ui/*.tsx`; check portal positioning, `useDeferredUnmount` hook order, `tw-animate-css` classes |
+| Streaming / SSE | `hooks/` streaming hook, `stores/` pending message queue |
+| Mention / file ref | `InputBar.mentions.ts`, `UserBubble.tsx` → `renderMentionSegments` |
+
+---
+
+## Tauri-aware patterns in the frontend
+
+Some behaviors differ between **web browser** and **Tauri webview**:
+
+- **Opening URLs externally** — in a browser use `window.open(url, '_blank')`; in Tauri use `@tauri-apps/plugin-opener` → `openUrl`. Detect with `window.__TAURI_INTERNALS__` or the `isTauri()` helper.
+- **File system access** — only available via Tauri commands, not the browser File API.
+- **Auth token injection** — desktop injects `X-Desktop-Token` via the Tauri sidecar handshake; web relies on cookie/session.
+
+---
+
+## Verification
+
+```bash
+cd web
+bun run typecheck
+bun run lint
+bun test --reporter=verbose
+```
+
+For visual sign-off after a component change, take a screenshot via `oad/frontend-debug`.

openagentd-1.78.0/.openagentd/skills/oad/debug/reference/tauri.md

@@ -0,0 +1,131 @@
+# Debug reference: Tauri / Rust (desktop & mobile)
+
+Use when the symptom is in the Rust layer — sidecar lifecycle, window management, tray, IPC, permissions, CSP, or native plugin behavior.
+
+---
+
+## Surface overview
+
+| Shell | Path | Role |
+|---|---|---|
+| Desktop | `desktop/src-tauri/` | Sidecar supervisor, tray, multi-window, auto-update |
+| Mobile | `mobile/src-tauri/` | Minimal shell, no sidecar, remote API only |
+
+---
+
+## Evidence commands
+
+### Fast Rust checks (no full Tauri build)
+
+```bash
+# Desktop
+cd desktop/src-tauri && cargo check
+cd desktop/src-tauri && cargo clippy -- -D warnings
+
+# Mobile
+cd mobile/src-tauri && cargo check
+cd mobile/src-tauri && cargo clippy -- -D warnings
+```
+
+### Run in dev mode (see live logs)
+
+```bash
+# Desktop — requires root `make dev` already running
+make -C desktop dev              # against external backend + Vite
+make -C desktop dev-bundled      # against bundled sidecar
+
+# Mobile
+make -C mobile dev               # Android
+make -C mobile ios-dev           # iOS simulator
+```
+
+Tauri dev mode prints Rust `log::` output to the terminal and writes to the platform log file. Filter with `RUST_LOG=debug` for verbose output.
+
+---
+
+## File map — desktop
+
+```
+desktop/src-tauri/
+  src/
+    main.rs              App entry, AppState, tray, window lifecycle, updater, IPC commands
+    sidecar.rs           Sidecar process supervisor, auth-token handshake
+  Cargo.toml             Deps: tauri 2, tokio, anyhow, serde, plugins
+  tauri.conf.json        Production config (updater endpoint, asset-protocol scope, bundles)
+  tauri.dev.conf.json    Dev against external backend + Vite
+  tauri.dev-bundled.conf.json  Dev against bundled sidecar
+  capabilities/          Tauri capability definitions (which commands/plugins frontend can call)
+  permissions/           Fine-grained permission sets
+  build.rs               Tauri build integration
+```
+
+## File map — mobile
+
+```
+mobile/src-tauri/
+  src/
+    main.rs              Minimal app entry — no sidecar, no tray
+  Cargo.toml             Minimal deps (no updater, no sidecar plugins)
+  tauri.conf.json        Window size (390×844 default), CSP, iOS/Android bundle config
+  build.rs               Tauri build integration
+```
+
+---
+
+## Common failure boundaries
+
+| Symptom | Where to look |
+|---|---|
+| Sidecar won't start / crashes | `src/sidecar.rs` — spawn args, env vars, port detection |
+| Auth token missing / rejected | `src/sidecar.rs` token generation, `core/desktop_auth.py` validation |
+| Window not appearing / wrong size | `src/main.rs` `WebviewWindowBuilder` + all three Tauri configs |
+| Tray icon / menu broken | `src/main.rs` `build_tray` / menu builders |
+| URL not opening in browser | `OpenerExt::open_url` — check `tauri-plugin-opener` is registered and permission granted |
+| IPC command not found | Missing `#[tauri::command]` + `invoke_handler` registration + capability entry |
+| CSP blocking a resource | `tauri.conf.json` → `app.security.csp` (update all config variants) |
+| Permission denied in webview | `capabilities/` — add the required permission set |
+| Auto-update not triggering | Updater endpoint URL, `pubkey`, `createUpdaterArtifacts: true` in production config |
+| iOS signing error | `mobile/ios-team.txt` (git-ignored) must contain valid Apple team ID |
+| Process not cleaned up on quit | `src/main.rs` `RunEvent::ExitRequested` / `WindowEvent::CloseRequested` handlers |
+
+---
+
+## Config variants — keep all three consistent (desktop)
+
+When changing window config, CSP, plugin config, or permissions, touch:
+1. `tauri.conf.json` — production
+2. `tauri.dev.conf.json` — external dev
+3. `tauri.dev-bundled.conf.json` — bundled dev
+
+---
+
+## Tauri ↔ frontend IPC patterns
+
+- **Rust → JS events**: `app_handle.emit("event-name", payload)` / `window.emit(...)`
+- **JS → Rust commands**: `invoke("command_name", args)` — requires `#[tauri::command]`, `invoke_handler`, and a `capabilities/` entry.
+- **Detect Tauri in frontend**: check `window.__TAURI_INTERNALS__` or use an `isTauri()` utility.
+
+---
+
+## Gotchas
+
+- `#[cfg(test)]` dialog stubs in `main.rs` — keep in sync with plugin API signatures.
+- Windows, macOS, and Linux have different process cleanup lifecycle events — test all three mentally before changing exit handlers.
+- Never commit `target/`, generated sidecar bundles, or machine-local `.openagentd/` state.
+- Mobile has **no sidecar** — always connects to an external API server.
+- `cargo check` is fast; `cargo tauri build` is slow — always verify with `check` first.
+
+---
+
+## Verification
+
+```bash
+cd desktop/src-tauri && cargo check && cargo clippy -- -D warnings
+cd mobile/src-tauri  && cargo check && cargo clippy -- -D warnings
+```
+
+For a full build smoke test:
+
+```bash
+make -C desktop build   # release bundle (slow — only when packaging)
+```

{openagentd-1.76.2 → openagentd-1.78.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openagentd
-Version: 1.76.2
+Version: 1.78.0
 Summary: Self-hosted AI agents — streaming chat, tool use, persistent memory, multi-agent teams
 Project-URL: Repository, https://github.com/lthoangg/openagentd
 Project-URL: Bug Tracker, https://github.com/lthoangg/openagentd/issues

{openagentd-1.76.2 → openagentd-1.78.0}/app/agent/errors.py

@@ -27,6 +27,31 @@ Hierarchy::
 
 from __future__ import annotations
 
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from pydantic import ValidationError
+
+
+def format_validation_error(exc: "ValidationError") -> str:
+    """Return a clean, single-line summary of a Pydantic ValidationError.
+
+    Strips the docs URL, type codes, and raw input values that Pydantic
+    appends by default — leaving only the field path and human message,
+    which is all an LLM (or a user) needs to self-correct.
+
+    Example output::
+
+        members: Input should be a valid list
+        a -> 0 -> value: Input should be a valid integer; b: Field required
+    """
+    return "; ".join(
+        f"{' -> '.join(str(loc) for loc in e['loc'])}: {e['msg']}"
+        if e.get("loc")
+        else e["msg"]
+        for e in exc.errors(include_url=False)
+    )
+
 
 class OpenAgentdError(Exception):
     """Base exception for all OpenAgentd domain errors."""

{openagentd-1.76.2 → openagentd-1.78.0}/app/agent/providers/googlegenai/schemas.py

@@ -13,7 +13,7 @@ class GeminiBaseModel(BaseModel):
 
 class FunctionCall(GeminiBaseModel):
     name: str
-    args: dict[str, Any]
+    args: dict[str, Any] = Field(default_factory=dict)
     id: str | None = None
 
 

{openagentd-1.76.2 → openagentd-1.78.0}/app/agent/tools/registry.py

@@ -78,7 +78,11 @@ from pydantic import BaseModel, ValidationError, create_model
 
 from loguru import logger
 
-from app.agent.errors import ToolArgumentError, ToolExecutionError
+from app.agent.errors import (
+    ToolArgumentError,
+    ToolExecutionError,
+    format_validation_error,
+)
 
 
 class InjectedArg:
@@ -259,7 +263,7 @@ class Tool:
             validated_model = self._model(**llm_kwargs)
         except ValidationError as exc:
             raise ToolArgumentError(
-                f"Invalid arguments for tool '{self.name}': {exc}"
+                f"Invalid arguments for tool '{self.name}': {format_validation_error(exc)}"
             ) from exc
         if self._model_param is not None:
             # The function wants the validated model as a single argument.

{openagentd-1.76.2 → openagentd-1.78.0}/app/api/routes/agents.py

@@ -30,6 +30,7 @@ from app.api.routes.settings import (
     _provider_is_configured,
     _provider_saved_overrides,
 )
+from app.api.schemas.base import _validation_detail
 from app.api.schemas.agents import (
     AgentDeleteResponse,
     AgentDetail,
@@ -147,10 +148,7 @@ def _parse_content(name: str, content: str) -> AgentConfig:
     try:
         return AgentConfig.model_validate(raw_meta)
     except ValidationError as exc:
-        errors = "; ".join(
-            f"{'.'.join(str(p) for p in e['loc'])}: {e['msg']}" for e in exc.errors()
-        )
-        raise ValueError(errors) from exc
+        raise ValueError(_validation_detail(exc)) from exc
 
 
 def _require_frontmatter_name(name: str, content: str) -> None:

{openagentd-1.76.2 → openagentd-1.78.0}/app/api/routes/team/files.py

@@ -590,7 +590,11 @@ async def get_coding_workspace_git_history(
             graph="",
         )
 
-    # 1. Fetch structured commits (limit + 1 to detect next page)
+    # 1. Fetch structured commits (limit + 1 to detect next page).
+    #    Fields: sha, short_sha, author_name, author_email, timestamp, subject,
+    #            body, refs — separated by NUL (\x00).  Each *record* ends with
+    #    the ASCII Record Separator (\x1e) so that multi-line body text doesn't
+    #    break parsing.
     log_args = ["log"]
     if cursor:
         log_args.extend([cursor, "--skip=1"])
@@ -598,7 +602,7 @@ async def get_coding_workspace_git_history(
         [
             "-n",
             str(limit + 1),
-            "--pretty=format:%H%x00%h%x00%an%x00%ae%x00%at%x00%s%x00%d",
+            "--pretty=format:%H%x00%h%x00%an%x00%ae%x00%at%x00%s%x00%b%x00%d%x1e",
         ]
     )
     if all_branches:
@@ -607,33 +611,36 @@ async def get_coding_workspace_git_history(
     commits_out = await _run_git(resolved, *log_args)
     commits = []
     if commits_out:
-        # Split by lines and parse each line
-        for line in commits_out.splitlines():
-            if not line:
+        # Split records on \x1e (each record ends with it, so the last split
+        # element will be an empty string — skip it).
+        for record in commits_out.split("\x1e"):
+            record = record.strip("\n")
+            if not record:
                 continue
-            parts = line.split("\x00")
-            if len(parts) >= 6:
-                try:
-                    timestamp = int(parts[4])
-                except ValueError:
-                    timestamp = 0
-
-                refs = (
-                    parts[6].strip(" ()")
-                    if len(parts) > 6 and parts[6].strip()
-                    else None
-                )
-                commits.append(
-                    GitCommit(
-                        sha=parts[0],
-                        short_sha=parts[1],
-                        author_name=parts[2],
-                        author_email=parts[3],
-                        timestamp=timestamp,
-                        subject=parts[5],
-                        refs=refs,
-                    )
+            parts = record.split("\x00")
+            if len(parts) < 6:
+                continue
+            try:
+                timestamp = int(parts[4])
+            except ValueError:
+                timestamp = 0
+
+            body_raw = parts[6].strip() if len(parts) > 6 else ""
+            refs = (
+                parts[7].strip(" ()\n") if len(parts) > 7 and parts[7].strip() else None
+            )
+            commits.append(
+                GitCommit(
+                    sha=parts[0],
+                    short_sha=parts[1],
+                    author_name=parts[2],
+                    author_email=parts[3],
+                    timestamp=timestamp,
+                    subject=parts[5],
+                    body=body_raw or None,
+                    refs=refs,
                 )
+            )
 
     # 2. Determine next_cursor and slice
     next_cursor = None

{openagentd-1.76.2 → openagentd-1.78.0}/app/api/schemas/base.py

@@ -21,4 +21,4 @@ class _ExcludeNoneModel(BaseModel):
 
 def _validation_detail(exc: ValidationError) -> str:
     """Extract a human-readable detail string from a Pydantic ValidationError."""
-    return "; ".join(e["msg"] for e in exc.errors())
+    return "; ".join(e["msg"] for e in exc.errors(include_url=False))

{openagentd-1.76.2 → openagentd-1.78.0}/app/api/schemas/team.py

@@ -84,6 +84,7 @@ class GitCommit(BaseModel):
     author_email: str
     timestamp: int
     subject: str
+    body: str | None = None
     refs: str | None = None
 
 

openagentd-1.78.0/app/version.txt

@@ -0,0 +1 @@
+1.78.0

{openagentd-1.76.2 → openagentd-1.78.0}/desktop/src-tauri/Cargo.lock

@@ -403,9 +403,9 @@ checksum = "72f5acc6cb2ba439de613abc23857ec3d78374d8ed5ac84e9d11336e87da8649"
 
 [[package]]
 name = "byte-unit"
-version = "5.2.4"
+version = "5.2.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ea91eef119026f16da52985896b3572d9abbb14f1ea5c9417fdf05da8309b66e"
+checksum = "4a813de7f2bbedb7dce265b64f1cf5908ebe4d56281ece8d847e98113788b9b0"
 dependencies = [
  "rust_decimal",
  "schemars 1.2.1",
@@ -483,9 +483,9 @@ dependencies = [
 
 [[package]]
 name = "camino"
-version = "1.2.3"
+version = "1.2.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b4ce8d3bd5823c7504d3f579f13e7b2f3da252fcb938c594d5680ee508bf846f"
+checksum = "5f2d30e4173c4026932d51d31d6b0613b1fd3014bf3f9f8943d4ba139c437ba0"
 dependencies = [
  "serde_core",
 ]
@@ -2603,7 +2603,7 @@ dependencies = [
 
 [[package]]
 name = "openagentd-desktop"
-version = "1.76.2"
+version = "1.78.0"
 dependencies = [
  "anyhow",
  "block2",

{openagentd-1.76.2 → openagentd-1.78.0}/desktop/src-tauri/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "openagentd-desktop"
-version = "1.76.2"
+version = "1.78.0"
 description = "OpenAgentd — on-machine AI agents (desktop shell)"
 authors = ["openagentd contributors"]
 license = "Apache-2.0"

{openagentd-1.76.2 → openagentd-1.78.0}/desktop/src-tauri/src/main.rs

@@ -177,28 +177,6 @@ fn configure_window_chrome(
     }
 }
 
-/// Re-assert the macOS overlay title-bar chrome on `window`.
-///
-/// New windows occasionally lose their traffic-light vertical centring after
-/// the webview triggers an AppKit relayout (e.g. navigating into
-/// cockpit/coding mode). Re-applying ``TitleBarStyle::Overlay`` toggles the
-/// window style mask (tao `set_style_mask_sync`), which forces AppKit to
-/// recompute the standard window buttons and marks the content view for
-/// redraw. tao's `draw_rect` then re-runs its stored `inset_traffic_lights`,
-/// restoring the configured 12/22 pt position. No-op off macOS.
-#[tauri::command]
-fn reapply_window_chrome(window: tauri::WebviewWindow) -> Result<(), String> {
-    #[cfg(target_os = "macos")]
-    {
-        use tauri::TitleBarStyle;
-        window
-            .set_title_bar_style(TitleBarStyle::Overlay)
-            .map_err(|e| format!("{e:#}"))?;
-    }
-    #[cfg(not(target_os = "macos"))]
-    let _ = window;
-    Ok(())
-}
 
 #[derive(Clone, Serialize)]
 struct BackendReady {
@@ -1205,7 +1183,17 @@ async fn run_update_install(app: AppHandle) -> Result<(), String> {
 
     update_tray_status(&app, "Status: Installing update…");
     log::info!("updater install applying version={}", update.version);
-    update.install(bytes).map_err(|e| {
+    // `tauri_plugin_updater`'s `install()` performs heavy filesystem work
+    // (decompression, bundle copy, codesign verification on macOS). Running it
+    // directly on an async task blocks the tokio thread, which can starve the
+    // logger and prevent the subsequent `run_on_main_thread` closure from being
+    // scheduled — causing `restart()` to never fire and the 60 s frontend
+    // timeout to trigger. Off-load to a dedicated blocking thread so the async
+    // runtime stays responsive throughout the install.
+    let install_result = tokio::task::spawn_blocking(move || update.install(bytes))
+        .await
+        .map_err(|e| format!("Install task panicked: {e}"))?;
+    install_result.map_err(|e| {
         update_tray_status(&app, "Status: Running");
         format!("Failed to install update: {e}")
     })?;
@@ -2190,7 +2178,6 @@ fn main() {
             app_use_bundled_backend,
             app_stop_bundled_backend,
             app_new_window,
-            reapply_window_chrome,
             set_tray_session,
             updater_check,
             updater_download,

{openagentd-1.76.2 → openagentd-1.78.0}/desktop/src-tauri/tauri.conf.json

@@ -2,7 +2,7 @@
   "$schema": "https://schema.tauri.app/config/2",
   "productName": "OpenAgentd",
   "mainBinaryName": "OpenAgentd",
-  "version": "1.76.2",
+  "version": "1.78.0",
   "identifier": "com.openagentd.desktop",
   "build": {
     "devUrl": "http://localhost:5173",