@inline-openclaw/inline 0.0.56 → 0.0.58

package/docs/create-inline-bot.md

@@ -0,0 +1,64 @@
+# Create an Inline Bot Token
+
+Use either the macOS app or the CLI to create a bot and get its token.
+
+## Option A: macOS App (Preferences)
+
+1. Open Inline for macOS.
+2. Open `Inline -> Settings…` (`Cmd+,`).
+3. Go to `Bots`.
+4. In `Create Bot`, enter:
+   - `Name`
+   - `Username` (must end with `bot`, case-insensitive)
+5. Click `Create Bot`.
+6. Copy the `New Token` shown after creation.
+
+Notes:
+- You can create up to 5 bots.
+- In `Your Bots`, you can reveal or rotate tokens later.
+
+## Option B: Inline CLI
+
+1. Login:
+
+```sh
+inline auth login
+```
+
+2. Create the bot:
+
+```sh
+inline bots create --name "My Inline Bot" --username myinlinebot
+```
+
+Optional: add the bot to a space on creation:
+
+```sh
+inline bots create --name "My Inline Bot" --username myinlinebot --add-to-space 123
+```
+
+3. Reveal token:
+
+```sh
+inline bots reveal-token --bot-user-id <BOT_USER_ID>
+```
+
+Tip: if you do not know the bot user id:
+
+```sh
+inline bots list
+```
+
+## Use Token in OpenClaw
+
+Set token under `channels.inline.token`:
+
+```yaml
+channels:
+  inline:
+    enabled: true
+    token: "<INLINE_BOT_TOKEN>"
+```
+
+Alternatively, set `INLINE_TOKEN` in the OpenClaw gateway environment.
+`INLINE_BOT_TOKEN` is also accepted as a compatibility alias.

package/docs/openclaw-setup.md

@@ -0,0 +1,138 @@
+# OpenClaw Setup Guide
+
+This guide is the fastest way to get the Inline channel working in OpenClaw.
+
+Need a bot token first? See `docs/create-inline-bot.md`.
+
+## 1) Install Plugin
+
+```sh
+openclaw plugins install @inline-openclaw/inline
+```
+
+If already installed, update to latest:
+
+```sh
+openclaw config set plugins.installs.inline.spec '"@inline-openclaw/inline@latest"'
+openclaw plugins update inline
+openclaw gateway restart
+openclaw plugins list
+openclaw channels status
+```
+
+After the first install, an OpenClaw owner can run `/inline_update` in Inline
+and then `/restart`; routine plugin updates do not require host shell access.
+
+## 2) Configure Inline Channel
+
+Minimal config (token-only):
+
+```yaml
+channels:
+  inline:
+    enabled: true
+    token: "<INLINE_BOT_TOKEN>"
+```
+
+Notes:
+- `baseUrl` defaults to `https://api.inline.chat`.
+- Instead of storing `token`, you can set `INLINE_TOKEN` in the gateway environment. `INLINE_BOT_TOKEN` is also accepted.
+- `defaultTo` is optional and gives outbound sends a fallback target, for example `chat:<id>` or `user:<id>`, when no explicit target is supplied.
+- If you add an explicit `plugins.entries` block, the plugin entry id is `inline`.
+- For multi-bubble replies, enable `channels.inline.blockStreaming: true`.
+- For reply-driven group flows, set `channels.inline.replyToBotWithoutMention: true`.
+- Groups use `groupPolicy: "open"` and `requireMention: true` by default, so any group can reach the channel but only an explicit mention wakes the bot unless configured otherwise.
+- Inline uses OpenClaw's group-history default; set `channels.inline.historyLimit` to override it, or use `messages.groupChat.historyLimit` as a global fallback.
+- Inline reply-thread handling is available by default, so OpenClaw `threadId` can map to real Inline reply-thread chats.
+- `replyToId` is still a message reply id. Inline reply-thread behavior does not replace ordinary message replies.
+- Legacy `capabilities.replyThreads` settings are ignored; use `replyThreadMode` to control automatic routing.
+- In an Inline group, authorized users can run `/threadreply` to choose this chat's automatic reply-thread mode with buttons.
+- Bot-participated reply threads and Inline dialogs already marked `FOLLOWING` continue without an explicit mention by default, matching Slack. Reply threads also persist recent participation so sparse follow-ups still route. Set `replyThreadRequireExplicitMention: true` globally, per account, or per group if a chat should require `@bot` on every thread message.
+- Reply-thread context defaults to nearby parent-chat messages, the anchor message, and child-thread history. Set `replyThreadParentHistoryLimit: 0` only when a chat should stay strictly thread-local.
+- Use `inline_parent_context` from a reply-thread session when the agent needs more complete parent-chat history than the automatic context window.
+- Inline current-message media is attached like native channels. Reply-thread anchor media is summarized as context and is not promoted to current-message media on every child-thread turn.
+- Message actions include reply/read/search/edit/reactions/channel and participant management; gate groups via `channels.inline.actions.*`.
+- Passive reaction notifications default to `channels.inline.reactionNotifications: "own"` for bot-authored messages. Set it to `"off"` to suppress queued reaction events, `"all"` to queue reactions on any authorized message, or `"allowlist"` with `reactionAllowlist` for selected reaction senders; named accounts can override the same fields.
+- Media uploads (image/video/document) are enabled by default for `mediaUrl` sends; set `channels.inline.mediaMaxMb` if you need a lower cap.
+- Native exec approvals use `channels.inline.execApprovals`. Set `approvers` to Inline user IDs such as `123` or `user:123`, or rely on numeric `commands.ownerAllowFrom`; `target` defaults to approver DMs.
+
+Access defaults:
+- DMs use `dmPolicy: "pairing"` unless configured.
+- For private bots, use `dmPolicy: "allowlist"` with your Inline user id in `allowFrom`.
+- For public/demo bots, use `dmPolicy: "open"` with `allowFrom: ["*"]`.
+- Groups use `groupPolicy: "open"` unless configured.
+- Group messages require a bot mention by default (`requireMention: true`). Use `groups` to override mention behavior for selected chats.
+- For selected groups, list numeric chat ids under `groups`.
+- Use `groupAllowFrom` for an account-wide group sender allowlist, or `groups.<chat>.allowFrom` for per-group sender allowlists.
+- `groupAllowFrom` is optional; use it only when specific senders inside allowed groups should be able to trigger the bot.
+
+Example reply-thread toggle:
+
+```yaml
+channels:
+  inline:
+    replyThreadRequireExplicitMention: false
+    replyThreadParentHistoryLimit: 10
+    groups:
+      "123":
+        replyThreadMode: "thread"
+        replyThreadRequireExplicitMention: true
+        replyThreadParentHistoryLimit: 2
+```
+
+## 3) Start Gateway
+
+Foreground:
+
+```sh
+openclaw gateway run
+```
+
+Service:
+
+```sh
+openclaw gateway start
+```
+
+If another gateway service is already running:
+
+```sh
+openclaw gateway stop
+openclaw gateway start
+```
+
+## 4) Verify Health
+
+```sh
+openclaw plugins list
+openclaw status --deep
+```
+
+Expected:
+- Plugin `inline` is loaded.
+- Channel `Inline` is configured and running/connected.
+
+## 5) Common Fixes
+
+- `Config validation failed: plugins.entries.inline: plugin not found: inline`
+  - Ensure the plugin is installed and discovered (`openclaw plugins list`).
+- `Inline: SETUP / no token`
+  - Ensure `channels.inline.token`, `INLINE_TOKEN`, or `INLINE_BOT_TOKEN` is set.
+  - Update plugin to latest (commands above).
+- `doctor --fix` suggests changing the Inline plugin entry
+  - Keep the plugin entry id as `inline`, then re-run `openclaw plugins list` and `openclaw channels status`.
+
+## Recommended Hardening (After Basic Setup)
+
+```sh
+openclaw config set session.dmScope '"per-channel-peer"'
+```
+
+Add plugin allowlist to reduce extension loading risk:
+
+```yaml
+plugins:
+  allow:
+    - inline
+    # add other plugin ids you trust
+```

package/openclaw.plugin.json

@@ -106,7 +106,7 @@
               "type": "object",
               "additionalProperties": false,
               "properties": {
-                "requireMention": { "type": "boolean" },
+                "requireMention": { "type": "boolean", "default": true },
                 "replyThreadMode": { "type": "string", "enum": ["auto", "thread", "main"] },
                 "replyThreadAutoCreateMinMessages": { "type": "integer", "minimum": 0 },
                 "replyThreadRequireExplicitMention": { "type": "boolean" },
@@ -141,7 +141,7 @@
               }
             }
           },
-          "requireMention": { "type": "boolean" },
+          "requireMention": { "type": "boolean", "default": true },
           "replyThreadMode": { "type": "string", "enum": ["auto", "thread", "main"] },
           "replyThreadAutoCreateMinMessages": { "type": "integer", "minimum": 0 },
           "replyThreadRequireExplicitMention": { "type": "boolean" },
@@ -462,7 +462,8 @@
                       "additionalProperties": false,
                       "properties": {
                         "requireMention": {
-                          "type": "boolean"
+                          "type": "boolean",
+                          "default": true
                         },
                         "replyThreadMode": {
                           "type": "string",
@@ -557,7 +558,8 @@
                     }
                   },
                   "requireMention": {
-                    "type": "boolean"
+                    "type": "boolean",
+                    "default": true
                   },
                   "replyThreadMode": {
                     "type": "string",
@@ -970,6 +972,14 @@
           "label": "Inline",
           "help": "Inline channel configuration for bot tokens, direct messages, group chats, and Inline-specific reply behavior."
         },
+        "name": {
+          "label": "Inline Account Name",
+          "help": "Optional display name for this Inline bot account in OpenClaw status and setup surfaces."
+        },
+        "enabled": {
+          "label": "Inline Account Enabled",
+          "help": "Enable or disable this Inline bot account without removing its configuration. Defaults to enabled."
+        },
         "baseUrl": {
           "label": "Inline API Base URL",
           "help": "Inline API base URL. Defaults to https://api.inline.chat."
@@ -1004,7 +1014,7 @@
         },
         "requireMention": {
           "label": "Inline Require Mention",
-          "help": "Require an explicit bot mention in group chats before routing a message to the agent."
+          "help": "Require an explicit bot mention in group chats before routing a message to the agent. Defaults to true."
         },
         "replyToBotWithoutMention": {
           "label": "Inline Reply Mention Bypass",
@@ -1014,6 +1024,10 @@
           "label": "Inline Reply Threads (Legacy)",
           "help": "Legacy compatibility flag. Inline reply-thread tools are available for bot accounts; use replyThreadMode to control automatic routing."
         },
+        "capabilities": {
+          "label": "Inline Capabilities (Legacy)",
+          "help": "Legacy compatibility settings. Inline reply-thread tools are available by default; use replyThreadMode for automatic routing."
+        },
         "replyThreadMode": {
           "label": "Inline Reply Thread Mode",
           "help": "Controls routing for Inline group replies that start from parent-chat messages: \"auto\" creates a child reply thread only when the user explicitly asks for one, \"thread\" routes every eligible top-level turn into a child reply thread anchored to the triggering message, and \"main\" keeps automatic replies in the parent chat. Explicit thread tools remain available."
@@ -1034,6 +1048,10 @@
           "label": "Inline Group Overrides",
           "help": "Per-group settings keyed by Inline chat ID, including sender allowlist, mention, reply-thread mode, reply-thread mention/history policy, deprecated threshold compatibility, and tool policy overrides."
         },
+        "systemPrompt": {
+          "label": "Inline System Prompt",
+          "help": "Additional system guidance applied to agent turns from this Inline account. Per-group systemPrompt values can override it."
+        },
         "accounts": {
           "label": "Inline Accounts",
           "help": "Named Inline bot accounts. Account entries accept the same token, access, group, reply-thread, command, and streaming fields as channels.inline."
@@ -1042,6 +1060,22 @@
           "label": "Inline Message Actions",
           "help": "Enable or disable Inline message-tool action groups such as send, reactions, read, search, translate, edit, participants, message pins, and permissions."
         },
+        "historyLimit": {
+          "label": "Inline Group History Limit",
+          "help": "Maximum recent group messages included as agent context. Uses the OpenClaw group-history default when unset."
+        },
+        "dmHistoryLimit": {
+          "label": "Inline DM History Limit",
+          "help": "Maximum recent direct-message messages included as agent context. Uses the OpenClaw DM-history default when unset."
+        },
+        "parseMarkdown": {
+          "label": "Inline Markdown",
+          "help": "Parse OpenClaw Markdown into Inline formatting for outgoing text. Defaults to true."
+        },
+        "mediaMaxMb": {
+          "label": "Inline Media Limit",
+          "help": "Maximum media size in megabytes accepted for Inline upload and send actions. Must be greater than zero."
+        },
         "reactionNotifications": {
           "label": "Inline Reaction Notifications",
           "help": "Controls passive reaction notifications queued into the agent session: \"off\", \"own\" for bot-authored messages, \"all\" for any authorized message, or \"allowlist\" for reactionAllowlist senders."
@@ -1062,6 +1096,10 @@
           "label": "Inline Bot Command Sync",
           "help": "Controls startup sync for Inline bot commands."
         },
+        "commands": {
+          "label": "Inline Bot Commands",
+          "help": "Configure startup synchronization of OpenClaw native and skill commands to the Inline bot command menu."
+        },
         "commands.nativeSkills": {
           "label": "Inline Skill Command Sync",
           "help": "Controls whether skill bot commands are included in startup command sync."
@@ -1094,6 +1132,34 @@
           "label": "Inline Streaming Mode",
           "help": "Unified Inline stream mode: \"off\" | \"partial\" | \"block\" | \"progress\". \"partial\" edits visible answer previews; \"progress\" sends a silent temporary progress message, edits it while OpenClaw works, and deletes it before the final reply."
         },
+        "textChunkLimit": {
+          "label": "Inline Text Chunk Limit",
+          "help": "Maximum characters per outgoing Inline text chunk. Must be greater than zero."
+        },
+        "chunkMode": {
+          "label": "Inline Chunk Mode",
+          "help": "Choose length-based or newline-aware splitting for outgoing Inline text chunks."
+        },
+        "streamMode": {
+          "label": "Inline Stream Mode (Legacy)",
+          "help": "Legacy alias for the Inline streaming mode. Prefer streaming.mode for new configuration."
+        },
+        "draftChunk": {
+          "label": "Inline Draft Chunking (Legacy)",
+          "help": "Legacy partial-preview chunk thresholds. Prefer streaming.preview.chunk for new configuration."
+        },
+        "blockStreaming": {
+          "label": "Inline Block Streaming (Legacy)",
+          "help": "Legacy switch for block-style streaming. Prefer streaming.mode=\"block\" and streaming.block.enabled."
+        },
+        "streamViaEditMessage": {
+          "label": "Inline Edit Streaming (Legacy)",
+          "help": "Legacy switch for editing one Inline message during partial streaming. Prefer streaming.mode=\"partial\"."
+        },
+        "blockStreamingCoalesce": {
+          "label": "Inline Block Coalescing (Legacy)",
+          "help": "Legacy block coalescing thresholds. Prefer streaming.block.coalesce for new configuration."
+        },
         "streaming.mode": {
           "label": "Inline Streaming Mode",
           "help": "Canonical Inline stream mode: \"off\" | \"partial\" | \"block\" | \"progress\". Use \"progress\" for the separate temporary working indicator; it is independent from final answer streaming."

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@inline-openclaw/inline",
-  "version": "0.0.56",
+  "version": "0.0.58",
   "license": "Apache-2.0",
   "repository": {
     "type": "git",
@@ -13,6 +13,8 @@
     "dist/*.js.map",
     "dist/index.d.ts",
     "dist/index.d.ts.map",
+    "README.md",
+    "docs/*.md",
     "LICENSE",
     "openclaw.plugin.json"
   ],
@@ -29,11 +31,11 @@
     "lint": "bunx oxlint --ignore-path ../.oxlintignore src vitest.config.ts",
     "prepack": "bun run build",
     "typecheck": "tsc -p tsconfig.json --noEmit",
-    "test": "vitest run --coverage",
+    "test": "vitest run --coverage --maxWorkers=2",
     "check": "bun run typecheck && bun run lint && bun run test && bun run build"
   },
   "dependencies": {
-    "@inline-chat/realtime-sdk": "0.0.14",
+    "@inline-chat/realtime-sdk": "0.0.15",
     "zod": "4.4.3"
   },
   "peerDependencies": {