@sellable/mcp 0.1.127 → 0.1.129

package/skills/create-campaign-v2/references/watch-guide-narration.md

@@ -35,6 +35,7 @@ stage labels. The driver label comes from the watch URL mode (`mode=claude` or
   Apollo, or existing list.
 - During search iteration, say what you tried, what you are trying next, what
   sample you are checking, and why that helps this campaign.
+- At filter choice, do not say filtering the batch before rubrics are saved.
 - Avoid internal terms: MCP, tool, currentStep, workflow table, scout, debug.
 - Do not invent time estimates.
 
@@ -81,7 +82,7 @@ Default source funnel:
   "stage": "find-leads",
   "headline": "Choosing the lead source",
   "visibleState": "The browser is on Pick Provider while Codex explains the source path before opening a provider lane.",
-  "agentIntent": "Codex will start with Signal Discovery when warm LinkedIn engagement looks plausible, then switch to Sales Nav or Prospeo only if the quick viability math fails.",
+  "agentIntent": "Codex will start with people engaging with relevant LinkedIn posts when that looks plausible, then switch to Sales Nav to find ICP people actively posting on LinkedIn, then search by titles, and use Prospeo for a broader account/contact path. Sales Nav with recent activity remains the first LinkedIn fallback.",
   "nextAction": "Review source",
   "safety": "No leads import until you approve the source."
 }
@@ -191,10 +192,10 @@ After review batch import:
 ```json
 {
   "stage": "fit-message",
-  "headline": "Added to Campaign",
-  "visibleState": "The first review batch is in the campaign.",
-  "agentIntent": "Codex is adding fit filters to keep every send qualified while drafting the first message in the background.",
-  "nextAction": "Review fit rules and message",
+  "headline": "Do you want filters?",
+  "visibleState": "The first review batch is in the campaign. The list is ready to decide whether extra filtering is worth it.",
+  "agentIntent": "Codex is asking whether you want to further filter these leads before message review. The Message Draft Builder may prepare a template in the background, but review waits until the filter path is settled.",
+  "nextAction": "Choose filters or skip",
   "safety": "Nothing enriches, validates, or sends until the message is ready and approved.",
   "workerStatuses": {
     "leadFitBuilder": "running",
@@ -204,8 +205,11 @@ After review batch import:
 ```
 
 Use this after the selected review rows are present and before filters are
-saved. This is add-filters intent, not active filtering. `enableICPFilters`
-only becomes true after active filter criteria are saved.
+saved. This is add-filters intent, not active filtering. Tell the user to
+Choose filters or skip, and do not say filtering the batch before rubrics are
+saved. When the user chooses filters, immediately persist
+`enableICPFilters: true` and move to `create-icp-rubric` so the watched app
+shows Filter Rules while Codex defines the rules in chat.
 
 Fit + message:
 
@@ -228,14 +232,27 @@ Review ready:
 ```json
 {
   "stage": "review-ready",
-  "headline": "Review fit and message",
-  "visibleState": "The fit rules and message sample are ready.",
-  "agentIntent": "Approve them or ask for changes in chat.",
+  "headline": "Choose message mode",
+  "visibleState": "The Messages step shows Use Template and AI Generated.",
+  "agentIntent": "Use Template is recommended. AI Generated is an explicit opt-out that cancels or ignores the background template draft.",
   "nextAction": "Validation",
   "safety": "Codex will only continue after you approve."
 }
 ```
 
+Messages waiting for template:
+
+```json
+{
+  "stage": "review-ready",
+  "headline": "Waiting for the template",
+  "visibleState": "Use Template is selected, but the Message Draft Builder is still finishing.",
+  "agentIntent": "Codex is waiting for the Message Draft Builder and will not ask for approval until the template can be reviewed.",
+  "nextAction": "Review template",
+  "safety": "Late template output must not overwrite campaign state if AI Generated was selected."
+}
+```
+
 Blocked/recovering:
 
 ```json

package/skills/create-rubric/SKILL.md

@@ -51,9 +51,13 @@ Create and save a comprehensive lead scoring rubric that:
 
 ## Rubric Validation + Results
 
-- `mcp__sellable__check_rubric` - Queue enrichment + scoring sample (default 7)
+- `mcp__sellable__get_rows_minimal` - Read bounded review-batch cell IDs
+- `mcp__sellable__queue_cells` - Queue only bounded `enrichCellId` values
 - `mcp__sellable__wait_for_rubric_results` - Poll pass-rate results
 
+`check_rubric` is legacy/recovery only. Do not use it in normal
+create-campaign-v2 runs.
+
 </tools>
 
 <rubric_structure>
@@ -117,7 +121,7 @@ Required fields:
 
 - If **Keep & apply**:
   - Call `update_campaign({ campaignId, enableICPFilters: true, currentStep: "apply-icp-rubric" })`
-  - Then proceed to Phase 5 (enrich + score a sample).
+  - Then proceed to Phase 5 (queue bounded enrichment cells + wait for score results).
 
 **If no campaignOfferId:**
 Ask: "Which campaign should I create the rubric for? You can share the campaign ID or describe the target audience."
@@ -203,7 +207,9 @@ update_campaign({
 
 This triggers the campaign sync and moves the user to the filter-leads step where they can see the campaign table with scoring columns.
 
-Then ask the user if we should kick off enrichment for a sample:
+For normal create-campaign-v2 runs, the caller should immediately queue only the
+bounded review-batch enrichment cells and wait for scores. For standalone
+rubric editing, ask the user if we should kick off enrichment for a sample:
 
 "ICP filtering enabled — you should see the leads table with scoring columns now. Want me to run enrichment on the first 5 leads to check pass rates?"
 
@@ -212,9 +218,14 @@ Then ask the user if we should kick off enrichment for a sample:
 
 ## Phase 5: Enrich + Score a Sample
 
-1. Call `check_rubric` with `maxProspects: 5` — finds the first 5 rows' enrich column cells and kicks off enrichment + ICP scoring via the batch queue API.
-2. Call `wait_for_rubric_results` with `targetCount: 5` — polls until ICP scores land.
-3. Summarize pass rate (`passed/completed`, percent) and recommend:
+1. Call `get_rows_minimal({ tableId, limit: 5 })` or use the caller-provided
+   bounded review-batch rows.
+2. Call `queue_cells({ tableId, cellIds })` with only pending/error
+   `enrichCellId` values. Do not pass `icpCellId` values or full-table cells;
+   the workflow cascade handles downstream ICP scoring after enrichment.
+3. Call `wait_for_rubric_results` with `targetCount: 5` and `minPassedCount: 1`
+   — polls until ICP scores land.
+4. Summarize pass rate (`passed/completed`, percent) and recommend:
    - Tighten filters if pass rate is too high (>80% — too loose).
    - Loosen filters if pass rate is too low (<20% — too strict).
    - Continue when results look healthy (30-70% is typical).