@mseep/obsidian-agent-client 0.10.6

package/docs/usage/sending-images.md

@@ -0,0 +1,64 @@
+# Sending Images and Files
+
+Attach images and files to your messages to provide context to the AI agent.
+
+## Attaching Files
+
+### Drag and Drop
+
+1. Drag files from Finder/Explorer onto the input area
+2. The input area highlights when you drag over it
+3. Thumbnails (images) or file icons (other files) appear below the text area
+
+### Paste from Clipboard
+
+1. Copy a file or image to your clipboard
+2. Focus the input field
+3. Paste with `Cmd/Ctrl + V`
+
+<p align="center">
+  <img src="/images/sending-images.webp" alt="Sending Images and Files" width="400" />
+</p>
+
+::: tip
+Embedding images in messages requires agent support. If the agent doesn't support images, image files are sent as file references instead (via drag and drop or paste from Finder/Explorer).
+:::
+
+## Managing Attachments
+
+Attached files appear as thumbnails or file icons below the text area.
+
+- **Remove an attachment**: Hover over it and click the **×** button
+- **Attachments are sent with your message**: When you send, all attached files are included
+
+<p align="center">
+  <img src="/images/remove-image.webp" alt="Remove attachment button" width="400" />
+</p>
+
+## Supported Image Formats
+
+| Format | MIME Type |
+|--------|-----------|
+| PNG | `image/png` |
+| JPEG | `image/jpeg` |
+| GIF | `image/gif` |
+| WebP | `image/webp` |
+
+Non-image files have no format restriction.
+
+::: info
+Non-image files are sent as file path references. How the agent processes the file depends on its capabilities and available tools.
+:::
+
+## Limits
+
+| Limit | Value |
+|-------|-------|
+| Maximum image size | 5 MB per image |
+| Maximum attachments | 10 per message (images + files combined) |
+
+::: info
+If you exceed these limits, a notification will inform you:
+- `[Agent Client] Image too large (max 5MB)`
+- `[Agent Client] Maximum 10 attachments allowed`
+:::

package/docs/usage/session-history.md

@@ -0,0 +1,91 @@
+# Session History
+
+Resume previous conversations or branch off from past sessions.
+
+## Agent Support
+
+Session history features are agent-specific. Not all agents support all features.
+
+## Opening Session History
+
+Click the **History** button (clock icon) in the chat header to open the session history modal.
+
+<p align="center">
+  <img src="/images/session-history-button.webp" alt="Session history button in chat header" />
+</p>
+
+## Available Actions
+
+Depending on the agent's capabilities, you can perform the following actions:
+
+| Action | Description |
+|--------|-------------|
+| **Edit title** | Rename the session from the history modal |
+| **Restore** | Resume the session where you left off |
+| **Fork** | Create a new branch from that point in the conversation |
+| **Delete** | Remove the session from history |
+
+::: tip
+Not all actions are available for every agent. The modal shows only the actions supported by your current agent.
+:::
+
+## Session Storage
+
+Sessions are saved automatically when you send messages. The plugin stores:
+
+- **Session metadata**: Title (derived from your first message), timestamps, and working directory
+- **Message history**: Full conversation including agent responses, tool calls, and plans
+
+### Where Sessions Are Stored
+
+Sessions are saved in two places:
+
+- **Plugin side**: Stored locally in Obsidian's data folder
+- **Agent side**: Managed by the agent
+
+## Restore vs Fork
+
+### Restore
+
+Restoring a session continues the existing conversation:
+
+1. The agent reconnects to the previous session
+2. Your conversation history is displayed
+3. New messages continue the same session
+
+Use restore when you want to **continue where you left off**.
+
+### Fork
+
+Forking creates a new session branching from a previous point:
+
+1. A new session is created with a copy of the conversation up to that point
+2. The original session remains unchanged
+3. New messages go to the forked session
+
+Use fork when you want to **explore a different direction** without affecting the original conversation.
+
+## Deleting Sessions
+
+To delete a session:
+
+1. Click the **Delete** button (trash icon) on the session
+2. Confirm the deletion in the dialog
+
+::: warning
+Deletion removes the session from the plugin's local storage only. The session still exists on the agent side.
+:::
+
+## Troubleshooting
+
+### "This agent does not support session restoration"
+
+The current agent doesn't provide session restore/fork capabilities. You can still view and delete locally saved sessions.
+
+### "Preparing agent..."
+
+The agent is still initializing. Wait a moment for the agent to become ready.
+
+### "No previous sessions"
+
+No sessions have been saved yet for the current agent and vault. Start a new conversation to create your first session.

package/docs/usage/slash-commands.md

@@ -0,0 +1,44 @@
+# Slash Commands
+
+Use slash commands to trigger actions provided by your current agent.
+
+## Agent Support
+
+Slash commands are agent-specific. Not all agents support slash commands.
+
+::: tip
+If the input field placeholder shows `/ for commands`, the current agent supports slash commands. If not, the agent does not support this feature.
+:::
+
+## Using Slash Commands
+
+1. Type `/` in the input field
+2. A dropdown appears showing available commands
+3. Select a command or continue typing to filter
+4. Press `Enter` to execute the command
+
+<p align="center">
+  <img src="/images/slash-commands-1.webp" alt="Slash commands dropdown" width="400" />
+</p>
+
+<p align="center">
+  <img src="/images/slash-commands-2.webp" alt="Filtering slash commands" width="400" />
+</p>
+
+## Available Commands
+
+Available commands are determined entirely by the agent—not by this plugin. Each agent provides its own set of commands.
+
+For example, Claude Code offers commands like `/compact`, `/init`, and `/review`, while other agents may have completely different commands or none at all.
+
+Refer to your agent's documentation for a full list of supported commands.
+
+## Command Arguments
+
+Some commands accept arguments:
+
+```
+/command keyword
+```
+
+Type the command followed by a space and your arguments.

package/esbuild.config.mjs

@@ -0,0 +1,49 @@
+import esbuild from "esbuild";
+import process from "process";
+import { builtinModules } from "node:module";
+
+const banner = `/*
+THIS IS A GENERATED/BUNDLED FILE BY ESBUILD
+if you want to view the source, please visit the github repository of this plugin
+*/
+`;
+
+const prod = process.argv[2] === "production";
+
+const context = await esbuild.context({
+	banner: {
+		js: banner,
+	},
+	entryPoints: ["src/main.ts"],
+	bundle: true,
+	external: [
+		"obsidian",
+		"electron",
+		"@codemirror/autocomplete",
+		"@codemirror/collab",
+		"@codemirror/commands",
+		"@codemirror/language",
+		"@codemirror/lint",
+		"@codemirror/search",
+		"@codemirror/state",
+		"@codemirror/view",
+		"@lezer/common",
+		"@lezer/highlight",
+		"@lezer/lr",
+		...builtinModules,
+	],
+	format: "cjs",
+	target: "es2018",
+	logLevel: "info",
+	sourcemap: prod ? false : "inline",
+	treeShaking: true,
+	outfile: "main.js",
+	minify: prod,
+});
+
+if (prod) {
+	await context.rebuild();
+	process.exit(0);
+} else {
+	await context.watch();
+}

package/eslint.config.mjs

@@ -0,0 +1,25 @@
+import tsparser from "@typescript-eslint/parser";
+import { defineConfig } from "eslint/config";
+import obsidianmd from "eslint-plugin-obsidianmd";
+import tseslint from "typescript-eslint";
+
+export default defineConfig([
+	{
+		ignores: ["node_modules/", "main.js", "docs/"],
+	},
+	...obsidianmd.configs.recommended,
+	...tseslint.configs.recommended,
+	{
+		files: ["**/*.ts", "**/*.tsx"],
+		languageOptions: {
+			parser: tsparser,
+			parserOptions: { project: "./tsconfig.json" },
+		},
+		rules: {
+			// Preserve existing rules
+			"@typescript-eslint/no-unused-vars": ["error", { args: "none" }],
+			"@typescript-eslint/ban-ts-comment": "off",
+			"@typescript-eslint/no-empty-function": "off",
+		},
+	},
+]);