bereach-openclaw 1.2.0 → 1.3.2

package/README.md

@@ -1,6 +1,6 @@
 # BeReach — OpenClaw Plugin
 
-LinkedIn outreach automation via [BeReach](https://berea.ch). Registers 33 in-process tools and auto-reply commands.
+LinkedIn outreach automation via [BeReach](https://berea.ch). Registers 60+ in-process tools and instant status commands.
 
 ## Install
 
@@ -137,11 +137,11 @@ They should match. Per [OpenClaw Plugin Manifest](https://docs.openclaw.ai/plugi
 
 ## Usage
 
-### Tools (33 registered)
+### Tools (60+ registered)
 
 All BeReach operations are available as agent tools — the agent uses them automatically based on your requests. No MCP needed; tools run in-process via the `bereach` SDK.
 
-### Auto-reply commands
+### Instant status commands
 
 These execute instantly without invoking the AI:
 
@@ -161,8 +161,8 @@ openclaw bereach status
 
 | Component | Description |
 | --- | --- |
-| `src/tools/` | 33 tool definitions generated from OpenAPI |
-| `src/commands/` | Auto-reply commands + CLI |
+| `src/tools/` | 60+ tool definitions generated from OpenAPI |
+| `src/commands/` | Instant status commands + CLI |
 | `skills/bereach/SKILL.md` | Main behavioral skill |
 | `skills/bereach/sub/lead-magnet.md` | Lead magnet workflow |
 | `skills/bereach/sdk-reference.md` | SDK method reference |

package/openclaw.plugin.json

@@ -1,8 +1,8 @@
 {
   "id": "bereach-openclaw",
   "name": "BeReach",
-  "version": "1.2.0",
-  "description": "LinkedIn outreach automation — 33 tools, auto-reply commands",
+  "version": "1.3.2",
+  "description": "LinkedIn outreach automation — 60+ tools, instant status commands",
   "configSchema": {
     "type": "object",
     "additionalProperties": false,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "bereach-openclaw",
-  "version": "1.2.0",
+  "version": "1.3.2",
   "description": "BeReach LinkedIn automation plugin for OpenClaw",
   "license": "AGPL-3.0",
   "exports": {
@@ -15,12 +15,12 @@
     ]
   },
   "dependencies": {
-    "bereach": "^1.2.3"
+    "bereach": "^1.3.2"
   },
   "devDependencies": {
-    "@types/node": "^22.19.3",
+    "@types/node": "^25.5.0",
     "tsx": "^4.21.0",
     "typescript": "^5.9.3",
-    "vitest": "^4.0.18"
+    "vitest": "^4.1.0"
   }
 }

package/skills/bereach/SKILL.md

@@ -1,10 +1,17 @@
 ---
 name: bereach
 description: "Automate LinkedIn outreach via BeReach (berea.ch). Use when: prospecting, engaging posts, scraping engagement, searching LinkedIn, managing inbox, running campaigns, managing invitations, analytics, company pages, Sales Navigator, content engagement, feed monitoring. Requires BEREACH_API_KEY."
-lastUpdatedAt: 1773274027
+lastUpdatedAt: 1773675510
 metadata: { "openclaw": { "requires": { "env": ["BEREACH_API_KEY"] }, "primaryEnv": "BEREACH_API_KEY" } }
 ---
 
+<!--
+  AUTO-GENERATED FILE — DO NOT EDIT
+  Source of truth: skills/ directory
+  Edit the source file, then run: pnpm build:plugins
+  Any direct edit to this file WILL be overwritten.
+-->
+
 # BeReach — LinkedIn Outreach Automation
 
 Automate LinkedIn prospection and engagement via BeReach.
@@ -17,11 +24,22 @@ Automate LinkedIn prospection and engagement via BeReach.
 
 | Sub-skill             | Keywords                                                                                                                  | URL                                                     | lastUpdatedAt |
 | --------------------- | ------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------- | ------------- |
-| Lead Magnet           | lead magnet, comment to DM, resource delivery, post giveaway, auto-accept invitations, scheduled run, recap, campaign stats, pause | sub/lead-magnet.md       | 1773267546    |
-| Lead Gen              | lead gen, find leads, search, qualify, ICP, pipeline, build list, scrape engagement, competitor, grow database, prospecting, hashtag, Sales Navigator         | sub/lead-gen.md          | 1773269560    |
-| Outreach              | outreach, connect, DM, message, follow up, sequence, connection request, reply, personalized outreach, campaign outreach, warming, follow, like post           | sub/outreach.md          | 1773269560    |
+| Lead Magnet           | lead magnet, comment to DM, resource delivery, post giveaway, auto-accept invitations, scheduled run, recap, campaign stats, pause | sub/lead-magnet.md       | 1773675510    |
+| Lead Gen              | lead gen, find leads, search, qualify, ICP, pipeline, build list, scrape engagement, competitor, grow database, prospecting, hashtag, Sales Navigator         | sub/lead-gen.md          | 1773663610    |
+| Outreach              | outreach, connect, DM, message, follow up, sequence, connection request, reply, personalized outreach, campaign outreach, warming, follow, like post           | sub/outreach.md          | 1773616752    |
 | OpenClaw Optimization | openclaw, model, opus, sonnet, haiku, config, SOUL.md, heartbeat, prompt caching, AI cost reduction, /model               | openclaw-optimization.md | 1773274027    |
-| SDK Reference         | sdk, method, parameter, signature, reference, api, script                                                                 | sdk-reference.md         | 1773269560    |
+| SDK Reference         | sdk, method, parameter, signature, reference, api, script                                                                 | sdk-reference.md         | 1773665416    |
+
+### Workspace Templates
+
+At install or upgrade, set up the workspace with these templates:
+
+| Template | Purpose | URL |
+| --- | --- | --- |
+| SOUL.md | Always-loaded tool listing + session startup + sub-skills routing table | workspace/soul-template.md |
+| AGENTS.md | Lean sub-agent safety net | workspace/agents-template.md |
+
+**Step 0 (mandatory at install/upgrade):** Check if SOUL.md contains `<!-- bereach-workspace-v`. If missing: fetch soul-template.md from the URL above (or via `GET /api/onboarding/soul`), write it into the user's SOUL.md between the markers. Set up the pipeline cron (every 15min) if not already configured. This step cannot be skipped.
 
 ### Supporting files
 
@@ -72,14 +90,15 @@ The SDK auto-reads `BEREACH_API_KEY` from the environment. NEVER hardcode tokens
 1. **Token** — read `BEREACH_API_KEY` from environment with no default/fallback value. If not set, stop immediately.
 2. **Read all sub-skills** — fetch every sub-skill listed above.
 3. **Context acquisition** — build a rich picture of the user before any workflow. This context feeds all sub-skills and makes the agent smarter from the start.
-   - **Profile + language** — `getLinkedInProfile` (0 credits). Detect language from location, headline, posts. This becomes the default for all generated content.
+   - **Profile + language** — `profile.get` (0 credits). Detect language from location, headline, posts. This becomes the default for all generated content.
    - **Recent posts** — `getPosts({ count: 10 })` (0 credits). What topics resonate, tone, offer, engagement patterns.
    - **Followers** — `getFollowers({ count: 50 })` (1 credit/page). Reveals audience type; followers are warm leads already.
    - **Website** — if the profile contains a website URL, fetch and extract: what they sell, who they target, positioning.
    - **Documents/data** — if the user has provided files (ICP docs, prospect lists, CRM exports, target account lists), read them. No format requirement.
    - **Existing campaigns** — `getStats` per known campaign (0 credits). What's running, what worked, conversions.
-   - **Lead-gen state** — `contacts.listCampaigns({ type: "lead_gen" })` (0 credits) to see existing lead-gen campaigns. `agentState.get("lead-gen")` (0 credits) to load previous state (angles, watchlist, learning data). If state exists, the agent resumes where it left off.
-   - **Analytics** — `linkedinAnalytics.getFollowerAnalytics()` (1 credit). Follower demographics and growth trends. `linkedinAnalytics.getProfileViews()` (1 credit). Who's viewing the profile — warm leads.
+   - **Memory discovery** — `agentState.list({ keysOnly: true })` (0 credits). Returns all stored state keys and timestamps. Use this to discover what the agent remembers from previous sessions — lead-gen state, outreach state, user profile, campaign configs. Load relevant keys with `agentState.get(key)`. Prefer this over hardcoded key names.
+   - **Lead-gen state** — `contacts.listCampaigns({ type: "lead_gen" })` (0 credits) to see existing lead-gen campaigns. If `agentState.list` returned a `lead-gen` key, load it with `agentState.get("lead-gen")` (0 credits) to resume where the agent left off.
+   - **Analytics** — `profile.getFollowerAnalytics()` (1 credit). Follower demographics and growth trends. `profile.getProfileViews()` (1 credit). Who's viewing the profile — warm leads.
    - **Company pages** — `companyPages.list()` (1 credit). If the user administers company pages, discover them for company-level posting and analytics.
    - **Competitors** — infer from industry, profile, and posts. Note competitor names and profile URLs for the engagement watchlist.
 4. **Welcome** — personalized welcome using profile data and context. Suggest a first action based on their profile, posts, and industry.
@@ -109,25 +128,31 @@ The SDK auto-reads `BEREACH_API_KEY` from the environment. NEVER hardcode tokens
 
 - NEVER hardcode tokens. Read from `process.env.BEREACH_API_KEY`. If not set, exit immediately.
 - NEVER invent a tool/method name. Check the resources table below or tool schemas.
+- Profile inputs: URL, URN, or bare vanity name (e.g. `john-doe`).
 - If unsure about a method: check the SDK Reference sub-skill.
 
 ### SDK resources and methods
 
 | Resource | Methods |
 | --- | --- |
-| `client.linkedinScrapers` | `collectLikes`, `collectComments`, `collectCommentReplies`, `collectPosts`, `visitProfile`, `visitCompany`, `collectHashtagPosts`, `collectSavedPosts` |
-| `client.linkedinActions` | `connectProfile`, `listInvitations`, `acceptInvitation`, `sendMessage`, `replyToComment`, `likeComment`, `publishPost`, `commentOnPost`, `likePost`, `unlikePost`, `unlikeComment`, `editComment`, `editPost`, `repostPost`, `savePost`, `unsavePost`, `followProfile`, `unfollowProfile`, `followCompany`, `unfollowCompany`, `declineInvitation`, `withdrawInvitation`, `listSentInvitations` |
-| `client.linkedinChat` | `listConversations`, `searchConversations`, `findConversation`, `getMessages`, `archiveConversation`, `unarchiveConversation`, `starConversation`, `unstarConversation`, `listArchivedConversations`, `listStarredConversations`, `markAllRead`, `markSeen`, `reactToMessage`, `unreactToMessage`, `getUnreadCount`, `sendTypingIndicator` |
-| `client.linkedinSearch` | `unifiedSearch`, `searchPosts`, `searchPeople`, `searchCompanies`, `searchJobs`, `searchByUrl`, `resolveParameters` |
-| `client.profile` | `getLinkedInProfile`, `refresh`, `getPosts`, `getFollowers`, `getLimits`, `getCredits`, `listAccounts`, `updateAccount`, `getActivity`, `revalidate` |
-| `client.campaigns` | `getStatus`, `syncActions`, `getStats` |
-| `client.contacts` | `createCampaign`, `listCampaigns`, `addContacts`, `listContacts`, `updateContact`, `searchContacts`, `getStats`, `logActivity`, `getActivities` |
-| `client.linkedinAnalytics` | `getFollowerAnalytics`, `getPostAnalytics`, `getProfileViews`, `getSearchAppearances` |
-| `client.linkedinFeed` | `getFeed` |
-| `client.salesNavigator` | `search`, `searchPeople`, `searchCompanies` |
-| `client.companyPages` | `list`, `getAnalytics`, `getPermissions`, `getPosts` |
-| `client.settings` | `getDmPolling`, `updateDmPolling` |
-| `client.agentState` | `get`, `set` |
+| `client.scrapers` | `collectPosts`, `visitProfile` |
+| `client.linkedinScrapers` | `collectLikes`, `collectComments`, `visitCompany`, `listSavedPosts`, `collectHashtagPosts` |
+| `client.linkedInScrapers` | `collectCommentReplies`, `getFeed` |
+| `client.linkedInSearch` | `search`, `searchCompanies`, `resolveParameters` |
+| `client.linkedinSearch` | `searchPosts`, `searchPeople`, `searchJobs`, `searchByUrl` |
+| `client.actions` | `connectProfile`, `sendMessage`, `listSentInvitations`, `followCompany` |
+| `client.linkedinActions` | `listInvitations`, `acceptInvitation`, `replyToComment`, `likePost`, `withdrawInvitation`, `followProfile`, `editPost`, `editComment`, `repostPost`, `unlikePost`, `unlikeComment`, `unsavePost`, `unfollowCompany` |
+| `client.linkedInActions` | `likeComment`, `publishPost`, `createComment`, `declineInvitation`, `unfollowProfile`, `savePost` |
+| `client.profile` | `get`, `listAccounts`, `updateAccount`, `refresh`, `getPosts`, `getFollowers`, `getLimits`, `getCredits`, `getProfileViews`, `getSearchAppearances`, `getPostAnalytics`, `getFollowerAnalytics` |
+| `client.companyPages` | `list`, `getPosts`, `getPermissions`, `getAnalytics` |
+| `client.linkedinChat` | `listInboxConversations`, `getMessages`, `markAllRead`, `listArchived`, `react`, `sendTypingIndicator`, `getUnreadCount` |
+| `client.chat` | `searchConversations`, `findConversation`, `archive`, `unreact` |
+| `client.linkedInChat` | `markSeen`, `star`, `unstar`, `listStarred`, `unarchive` |
+| `client.campaigns` | `getStatus`, `sync`, `getStats` |
+| `client.contacts` | `upsert` |
+| `client.salesNavigatorSearch` | `search`, `searchCompanies` |
+| `client.salesNav` | `searchPeople` |
+| **API-only (REST)** | `contacts.*` (campaigns, addContacts, updateContact, searchContacts, etc.), `agentState.*`, `settings.*` — use tools or REST API, not SDK |
 
 ## Tone
 
@@ -139,10 +164,10 @@ These are technical constraints BeReach requires. Everything else, adapt as need
 
 1. **Dedup** — pass `campaignSlug` on every action. BeReach deduplicates by target automatically. Duplicates return `duplicate: true` and cost nothing. Pre-check: use `client.campaigns.getStatus()` or the `getStatus` tool.
 2. **Pacing** — after every SDK/tool call, sleep a random delay. Write actions (DM, reply, like, comment, connect, accept, follow, repost, edit, sync): `random delay 8-12s`. Read actions (visit, scrape, search, find, feed, analytics, count-0 checks): `random delay 2-6s`. Inbox management (archive, star, mark-read, react, typing): `random delay 1-3s`. On 429: wait the number of seconds from the error response, retry (max 3). If daily/weekly cap hit, switch action type.
-3. **Save incrementally** — persist tracking after each action, not at the end.
+3. **Save incrementally** — persist pipeline progress and tracking state after each action, not at the end. Contact data is auto-saved by API calls; only save operational state (e.g. which posts were scraped, discovery sources processed).
 4. **Limits check** — call `getLimits` once per day at session start.
 5. **Visit before connecting** — looks natural to LinkedIn.
-6. **Credits** — check with `getCredits` → `{credits: {current, limit, remaining, percentage}}`.
+6. **Credits** — check with `getCredits` → `{credits: {current, limit, remaining, percentage, isUnlimited}}`. When `isUnlimited` is true, `remaining` and `limit` are null — credits are unlimited, skip credit budgeting.
 7. **Connection requests are scarce** — only 30/day. Before sending one, check `pendingConnection` from the visit response: `"pending"` = already sent, skip. `"failed"` = failed today, skip. `"none"` = safe to send.
 8. **Error handling** — wrap every SDK call in a shared helper. The wrapper must check the **error type** to branch:
    - `429`: wait `error.retryAfter` seconds, retry (max 3). All retries fail → skip profile, move on.
@@ -153,47 +178,152 @@ These are technical constraints BeReach requires. Everything else, adapt as need
    - Parse the response body inside a try/catch — malformed JSON must not crash the script.
    - Never treat a failed call as "success with empty result".
    - Never `process.exit()` on transient errors — skip and continue.
+9. **Stage transitions** — lifecycle stages are forward-only: `contact` → `lead` → `qualified` → `approved`. The only exception is `rejected`, which can be set from any stage. Downgrade attempts are silently ignored by the API.
+
+## Platform Infrastructure
+
+These systems run automatically. The agent should rely on them, not replicate their work.
+
+### Auto-Upsert from Activity
+
+When the agent scrapes posts, reads inbox, or finds a conversation, BeReach auto-creates contacts in the CRM for every LinkedIn profile encountered. The agent does NOT need to manually upsert after every scrape.
+
+The agent should still:
+
+- Update `lifecycleStage`, `hotScore`, `qualificationNotes` on contacts it qualifies
+- Add contacts to campaigns with `addContacts`
+- Log activities for meaningful observations
+
+But basic contact creation (name + linkedinUrl) happens automatically.
+
+### Smart Contact Deduplication
+
+Contacts are auto-deduplicated across all identifier types: profile URL, vanity name (`publicIdentifier`), and LinkedIn URN (`profileUrn`). Duplicates are merged automatically — activities, campaign memberships, and profile data are preserved.
+
+The agent can use any identifier type and trust that the CRM resolves to the right contact.
+
+### Campaign Context
+
+Each campaign carries a `context` field — the agent's playbook in markdown. `listCampaigns()` returns it for every campaign. Use it to tailor messaging tone, ICP criteria, and follow-up cadence per campaign. Update with `contacts.updateCampaign({ campaignId, context: "..." })`.
 
 ## Workflows
 
 Each workflow is detailed in its sub-skill. Load the relevant sub-skill when needed.
 
-- **Lead Magnet** — deliver a resource to everyone who engages with a post (comments, likes, invitations). Multi-campaign config at `~/.bereach/lead-magnet.json`. → Lead Magnet sub-skill
-- **Lead Gen** — autonomous lead generation. Find and qualify leads daily through multi-angle search, engagement scraping, and a learning loop. BeReach campaigns as the contacts DB. State at `~/.bereach/lead-gen-state.json`. → Lead Gen sub-skill
+- **Lead Magnet** — deliver a resource to everyone who engages with a post (comments, likes, invitations). Config via `agentState.get("lead-magnet-config")`. → Lead Magnet sub-skill
+- **Lead Gen** — autonomous lead generation. Find and qualify leads daily through multi-angle search, engagement scraping, and a learning loop. BeReach campaigns as the contacts DB. State via `agentState.get("lead-gen")`. → Lead Gen sub-skill
 - **Outreach** — personalized LinkedIn outreach. Connect, message, follow up, adapt. Uses emergent planning (no DAG workflows). State via `agentState.get("outreach")`. → Outreach sub-skill
 
 ### Cron
 
 Crons are OpenClaw scheduled tasks. Create entries in `~/.openclaw/cron/jobs.json` (via `openclaw cron add`).
 
-**Lead Magnet** — 3 global crons. Campaign config at `~/.bereach/lead-magnet.json`. Each cron runs one script that loops over all enabled campaigns:
+All crons use `sessionTarget: "isolated"` with `payload.kind: "agentTurn"` — each spawns an isolated agent that runs independently and announces results when done. Delivery must include `channel` and `bestEffort: true`. Never use `exec()` directly from a cron; it blocks until the script finishes.
+
+**Lead Magnet** — 3 independent hourly crons. Campaign config at `agentState.get("lead-magnet-config")`. Each cron runs one script that loops over all enabled campaigns. They run in parallel since OpenClaw spawns them independently:
 
 ```json
-{ "id": "lm-comments", "every": "1h", "skill": "bereach", "sessionTarget": "spawn", "prompt": "Run lm-comments.ts and report recap" }
-{ "id": "lm-invitations", "every": "1h", "skill": "bereach", "sessionTarget": "spawn", "prompt": "Run lm-invitations.ts and report recap" }
-{ "id": "lm-connections", "every": "1h", "skill": "bereach", "sessionTarget": "spawn", "prompt": "Run lm-connections.ts and report recap" }
+{
+  "name": "Lead Magnet — Comments",
+  "enabled": true,
+  "schedule": { "kind": "cron", "expr": "0 * * * *", "tz": "Europe/Paris" },
+  "sessionTarget": "isolated",
+  "wakeMode": "now",
+  "payload": {
+    "kind": "agentTurn",
+    "message": "Run lm-comments.ts for all enabled campaigns. Load config via agentState, validate, process new commenters, report recap.",
+    "timeoutSeconds": 600
+  },
+  "delivery": { "mode": "announce", "channel": "webchat", "bestEffort": true }
+}
 ```
 
-**Lead Gen** — 1 daily cron. The spawned agent uses tools directly (no pre-generated script):
+```json
+{
+  "name": "Lead Magnet — Invitations",
+  "enabled": true,
+  "schedule": { "kind": "cron", "expr": "0 * * * *", "tz": "Europe/Paris" },
+  "sessionTarget": "isolated",
+  "wakeMode": "now",
+  "payload": {
+    "kind": "agentTurn",
+    "message": "Run lm-invitations.ts for all enabled campaigns. Load config via agentState, validate, accept invitations and deliver resource, report recap.",
+    "timeoutSeconds": 600
+  },
+  "delivery": { "mode": "announce", "channel": "webchat", "bestEffort": true }
+}
+```
 
 ```json
-{ "id": "lead-gen", "every": "1d", "skill": "bereach", "sessionTarget": "spawn", "prompt": "Run daily lead gen using sub/lead-gen.md playbook. Load state via agentState.get('lead-gen'), run priority channels, qualify, store, report recap." }
+{
+  "name": "Lead Magnet — Connections",
+  "enabled": true,
+  "schedule": { "kind": "cron", "expr": "0 * * * *", "tz": "Europe/Paris" },
+  "sessionTarget": "isolated",
+  "wakeMode": "now",
+  "payload": {
+    "kind": "agentTurn",
+    "message": "Run lm-connections.ts for all enabled campaigns. Load config via agentState, validate, check pending distance 2+ connections, report recap.",
+    "timeoutSeconds": 600
+  },
+  "delivery": { "mode": "announce", "channel": "webchat", "bestEffort": true }
+}
 ```
 
-**Outreach** — 1 daily cron. Uses emergent planning (no workflows):
+**Lead Gen** — 1 daily cron. The agent uses tools directly (no pre-generated script):
 
 ```json
-{ "id": "outreach", "every": "1d", "skill": "bereach", "sessionTarget": "spawn", "prompt": "Run daily outreach using sub/outreach.md playbook. Load state via agentState.get('outreach'). Process replies first, then accepted connections, then follow-ups, then new connections. Respect daily limits. Update state and recap." }
+{
+  "name": "Lead Gen — Daily",
+  "enabled": true,
+  "schedule": { "kind": "cron", "expr": "0 8 * * *", "tz": "Europe/Paris" },
+  "sessionTarget": "isolated",
+  "wakeMode": "now",
+  "payload": {
+    "kind": "agentTurn",
+    "message": "Run daily lead gen using sub/lead-gen.md playbook. Discover state via agentState.list({ keysOnly: true }), load 'lead-gen' key if present. Run priority channels, qualify, store, report recap.",
+    "timeoutSeconds": 900
+  },
+  "delivery": { "mode": "announce", "channel": "webchat", "bestEffort": true }
+}
 ```
 
-Always use `"sessionTarget": "spawn"` for crons that execute scripts or tool-based workflows. This spawns a sub-agent in an isolated session — it runs independently and pushes results back to the user when done, without blocking the main session. Never use `exec()` directly from a cron prompt; it blocks until the script finishes.
+**Outreach** — 1 daily cron. Uses emergent planning (no workflows):
+
+```json
+{
+  "name": "Outreach — Daily",
+  "enabled": true,
+  "schedule": { "kind": "cron", "expr": "0 9 * * *", "tz": "Europe/Paris" },
+  "sessionTarget": "isolated",
+  "wakeMode": "now",
+  "payload": {
+    "kind": "agentTurn",
+    "message": "Run daily outreach using sub/outreach.md playbook. Discover state via agentState.list({ keysOnly: true }), load 'outreach' key. Load campaign contexts via listCampaigns(). Process replies first, then accepted connections, then follow-ups, then new connections. Respect daily limits. Update state and recap.",
+    "timeoutSeconds": 900
+  },
+  "delivery": { "mode": "announce", "channel": "webchat", "bestEffort": true }
+}
+```
 
 **Update check** — detect-only, never auto-upgrade:
 
 ```json
-{ "id": "bereach-update-check", "every": "1d", "skill": "bereach", "sessionTarget": "spawn", "prompt": "Check if a new BeReach plugin version is available. Compare local version from package.json with `npm view bereach-openclaw version`. If newer: notify user once: 'BeReach vX.Y.Z is available (current: vA.B.C). Say upgrade bereach when ready.' Do NOT upgrade automatically." }
+{
+  "name": "BeReach — Update Check",
+  "enabled": true,
+  "schedule": { "kind": "cron", "expr": "0 10 * * *", "tz": "Europe/Paris" },
+  "sessionTarget": "isolated",
+  "wakeMode": "now",
+  "payload": {
+    "kind": "agentTurn",
+    "message": "Check if a new BeReach plugin version is available. Compare local version from package.json with `npm view bereach-openclaw version`. If newer: notify user once: 'BeReach vX.Y.Z is available (current: vA.B.C). Say upgrade bereach when ready.' Do NOT upgrade automatically.",
+    "timeoutSeconds": 120
+  },
+  "delivery": { "mode": "announce", "channel": "webchat", "bestEffort": true }
+}
 ```
 
 When the user says "upgrade bereach": follow the upgrade workflow in `upgrade-openclaw-bereach.md` — save user state, uninstall, reinstall, recreate scripts from new skill, wait for user confirmation before launching crons.
 
-Remove with `openclaw cron remove <id>`.
+Remove with `openclaw cron remove <name>`.

package/skills/bereach/openclaw-optimization.md

@@ -4,6 +4,13 @@ description: "Optimize OpenClaw for cost efficiency with Anthropic Claude. Use w
 lastUpdatedAt: 1773274027
 ---
 
+<!--
+  AUTO-GENERATED FILE — DO NOT EDIT
+  Source of truth: skills/ directory
+  Edit the source file, then run: pnpm build:plugins
+  Any direct edit to this file WILL be overwritten.
+-->
+
 # OpenClaw Optimization
 
 > **Anthropic only.** Skip this skill if the user is not on Anthropic Claude. Proposed during BeReach onboarding or on explicit request. If declined, don't bring it up again.