@sellable/install 0.1.196 → 0.1.198

package/bin/sellable-install.mjs

@@ -620,6 +620,9 @@ const CREATE_CAMPAIGN_ALLOWED_TOOLS = [
   "mcp__sellable__select_campaign_cells",
   "mcp__sellable__queue_campaign_cells",
   "mcp__sellable__wait_for_campaign_processing",
+  "mcp__sellable__start_prepare_campaign_messages",
+  "mcp__sellable__get_prepare_campaign_messages_status",
+  "mcp__sellable__cancel_prepare_campaign_messages",
   "mcp__sellable__revise_message_template_and_rerun",
   "mcp__sellable__update_campaign_brief",
   "mcp__sellable__update_campaign",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@sellable/install",
-  "version": "0.1.196",
+  "version": "0.1.198",
   "type": "module",
   "description": "One-command installer for Sellable MCP in Claude Code and Codex",
   "bin": {

package/skill-templates/create-ab-test.md

@@ -1,14 +1,24 @@
 ---
 name: create-ab-test
-description: Create a Sellable campaign A/B test from a clean source lead list.
+description: Create a Sellable campaign A/B test by duplicating a full Campaign A.
 visibility: public
 allowed-tools:
   - mcp__sellable__get_auth_status
+  - mcp__sellable__bootstrap_create_campaign
+  - mcp__sellable__get_subskill_prompt
+  - mcp__sellable__get_subskill_asset
   - mcp__sellable__get_active_workspace
+  - mcp__sellable__fetch_linkedin_profile
+  - mcp__sellable__complete_sender_research
+  - mcp__sellable__create_campaign
+  - mcp__sellable__update_campaign_brief
+  - mcp__sellable__update_campaign
+  - mcp__sellable__save_rubrics
   - mcp__sellable__get_campaign
   - mcp__sellable__get_campaign_context
   - mcp__sellable__get_campaign_navigation_state
   - mcp__sellable__get_campaign_messages_preview
+  - mcp__sellable__duplicate_campaign
   - mcp__sellable__prepare_campaign_ab_test
   - mcp__sellable__load_csv_linkedin_leads
   - mcp__sellable__wait_for_lead_list_ready
@@ -24,25 +34,65 @@ campaign-message approaches from the same source list.
 
 ## Opening Contract
 
-Start by identifying the source campaign, the one variable being tested, and
-the clean source lead origin. Then prepare the split and stop for review. The
-goal is two review-copy campaigns with clean split source lists, not a launched
-campaign.
+Start by identifying the campaign idea and the one variable being tested.
+Campaign A must be a real, fully prepared Sellable campaign before this A/B
+helper runs. If the user only supplied an idea, first run the normal
+`$sellable:create-campaign` flow for Campaign A: research the sender/company,
+create the campaign shell, build and approve the campaign brief, source or
+attach leads when appropriate, save any approved ICP filters, and get the
+message/template state into the campaign brief. Do not call
+`prepare_campaign_ab_test` with only `campaignName`/`campaignBrief`.
+
+Once Campaign A exists, the A/B step duplicates Campaign A through Sellable's
+product duplicate path and applies only the B variant delta. This preserves the
+approved brief, ICP filters, message setup, rubrics, and campaign-table
+structure so the test differs only by the chosen variable and, if approved, the
+split lead list.
 
 ## Required Flow
 
-1. Confirm the source campaign ID or resolve it with `get_campaign`.
+1. Decide the source mode:
+   - Existing Campaign A: confirm the source campaign ID or resolve it with
+     `get_campaign`.
+   - Idea-only request: run the full `$sellable:create-campaign` workflow first,
+     then use the created campaign ID as `sourceCampaignId`.
 2. Confirm the A/B variable in plain language. Keep variant B as the explicit
    change and leave variant A as the control unless the user names an A change.
-3. Verify the campaign has a clean source lead list. A clean source list is a
-   lead-list table, not the generated campaign workflow table.
-4. Call `prepare_campaign_ab_test` with `dryRun: true` first.
-5. Review split counts, duplicate/skipped counts, campaign names, and variant
-   brief deltas with the user.
-6. Only after review, call `prepare_campaign_ab_test` without `dryRun` using the
+3. Inspect Campaign A with `get_campaign` and `get_campaign_context`. Confirm
+   the brief exists and includes the approved messaging/template state when the
+   campaign has reached message review.
+4. If Campaign A has a clean `selectedLeadListId`, ask the user whether to split
+   Campaign A's lead list for this A/B test:
+   - Split the lead list: pass that clean list as `sourceLeadListId`.
+   - Do not split yet: omit `sourceLeadListId`; Campaign B is a duplicate for
+     review and no rows are imported.
+   Never silently split a list just because Campaign A has one.
+5. If a source lead list is supplied, verify it is clean. A clean source list is
+   a lead-list or signal-lead-list table, not the generated campaign workflow
+   table.
+6. Call `prepare_campaign_ab_test` with `dryRun: true` first.
+7. Review split counts when applicable, campaign names, and variant brief deltas
+   with the user.
+8. Only after review, call `prepare_campaign_ab_test` without `dryRun` using the
    same `idempotencyKey` if one was returned or supplied.
-7. Return the A and B campaign IDs, split lead-list IDs, counts, and the
-   explicit note that both campaigns are `not_started`.
+9. Return Campaign A, duplicated Campaign B, split lead-list IDs when supplied,
+   counts, and the explicit note that both campaigns are `not_started`.
+
+## Tool Inputs
+
+Use one of these shapes:
+
+- No split: pass `sourceCampaignId`, `variantName`, `variantBriefDelta`, and
+  any optional A-variant fields. The helper treats the source campaign as A and
+  duplicates it for B.
+- Approved split: pass `sourceCampaignId`, the approved clean
+  `sourceLeadListId`, `variantName`, `variantBriefDelta`, and any optional split
+  or A-variant fields. The helper creates deterministic A/B split lead lists and
+  review-copy campaigns from Campaign A.
+
+Do not pass `campaignName`, `campaignBrief`, `offerPositioning`,
+`clientProspectId`, or `senderLinkedinUrl` to `prepare_campaign_ab_test`.
+Those belong to the full create-campaign workflow that creates Campaign A.
 
 ## Anti-Patterns
 
@@ -60,18 +110,19 @@ campaign.
 If the source campaign has no clean source lead list, or the source list is
 polluted with workflow/output columns, ask for the original CSV or clean source
 list. Use `load_csv_linkedin_leads`; it strips Sellable workflow columns if a
-contaminated CSV is supplied, then creates a clean lead list. After the clean
-source is confirmed, retry `prepare_campaign_ab_test`.
+contaminated CSV is supplied, then creates a clean lead list. If no clean source
+is available, omit `sourceLeadListId` and create only the B duplicate for review.
 
 ## Output Contract
 
 Report:
 
-- source campaign ID
-- source lead-list ID
-- split strategy and counts
-- duplicate/skipped lead counts
-- A review campaign ID and split lead-list ID
-- B review campaign ID and split lead-list ID
+- source Campaign A ID
+- source lead-list ID when supplied
+- split strategy and counts when a source lead list was supplied
+- duplicate/skipped lead counts when a source lead list was supplied
+- Campaign A ID, or A review-copy campaign ID when a split was approved
+- duplicated Campaign B ID
+- split lead-list IDs when supplied
 - variant difference
 - `launchState: not_started` for both campaigns

package/skill-templates/create-campaign.md

@@ -46,6 +46,9 @@ allowed-tools:
   - mcp__sellable__select_campaign_cells
   - mcp__sellable__queue_campaign_cells
   - mcp__sellable__wait_for_campaign_processing
+  - mcp__sellable__start_prepare_campaign_messages
+  - mcp__sellable__get_prepare_campaign_messages_status
+  - mcp__sellable__cancel_prepare_campaign_messages
   - mcp__sellable__revise_message_template_and_rerun
   - mcp__sellable__update_campaign_brief
   - mcp__sellable__update_campaign
@@ -89,6 +92,18 @@ When filters are chosen, save rubrics, get filter approval, then wait for
 message-template approval before enrichment/filtering or Generate Message cells.
 After filter approval, Filter Leads should show `Filters saved + waiting for
 message approval` until the template is approved.
+After message approval, use `start_prepare_campaign_messages` for target-ready
+message preparation such as "prepare 100 messages, checking up to 300 rows".
+Treat `campaignId` as `CampaignOffer.id`, keep the default
+`approvalMode:"mark_ready"` unless the user explicitly asks to flip Approved
+cells, and poll `get_prepare_campaign_messages_status` for checked rows,
+prepared/approved count, target, row budget remaining, and stop reason. Low
+level selector/queue tools are diagnostics and recovery only for this lane. If
+the user asks to stop preparation, the target is wrong, or status shows the
+wrong campaign/table, call `cancel_prepare_campaign_messages`; otherwise do not
+cancel a healthy prepare run.
+Never call `start_campaign` from Prepare Messages; final launch remains a
+separate explicit user greenlight.
 Use Template is the default message path; AI Generated is only an explicit
 opt-out.
 
@@ -955,7 +970,8 @@ updates.
    observation state,
    `auto-execute-messaging` after at least one row passes and initial campaign-row
    messages are being generated or reviewed, `awaiting-user-greenlight` only
-   after generated campaign-row messages are approved, `settings` for sender
+   after generated campaign-row messages are approved and the Prepare Messages
+   job has reported compact checked/prepared/stop status, `settings` for sender
    selection, `sequence` after sender attach, and `send` once the recommended
    sequence is attached. Do not advance the step backward.
 7. Keep `selectedLeadListId` as the source list and `workflowTableId` as the