@sellable/install 0.1.104 → 0.1.106

package/README.md

@@ -61,11 +61,14 @@ Use the same public entrypoints in both hosts:
 
 - Claude Code: `/sellable:create-campaign`
 - Claude Code: `/sellable:interview`
+- Claude Code: `/sellable:load-voice`
 - Codex: `$sellable:create-campaign`
 - Codex: `$sellable:interview`
+- Codex: `$sellable:load-voice`
 - Codex Desktop plugin: `sellable@sellable`
 - Codex visible skill: `Sellable Create Campaign`
 - Codex visible skill: `Sellable Identity Interview`
+- Codex visible skill: `Sellable Load Voice`
 - Internal MCP workflow prompt: `create-campaign-v2`
 
 Do not ask users to run `/sellable:create-campaign-v2`,

package/agents/post-find-leads-message-scout.md

@@ -136,3 +136,47 @@ Return a concise status with:
 - one rendered omit/fallback sample
 - quality-gate pass/fail summary
 - whether final template review is ready or needs revision
+
+When the parent will show the recommendation in chat, format the customer-facing
+message review as Markdown with distinct copy blocks:
+
+````markdown
+## Message Template
+
+**Subject**
+
+```text
+{{tokenized_subject}}
+```
+
+**Body**
+
+```text
+{{tokenized_message_body}}
+```
+
+## Rendered Examples
+
+### Good token fill
+
+```text
+Subject: ...
+
+Hey First,
+
+...
+```
+
+### Good omit / fallback
+
+```text
+Subject: ...
+
+Hey First,
+
+...
+```
+````
+
+Use a table for token rules. Keep explanations outside the fenced `text` blocks
+so the user can quickly see exactly what copy is being approved.

package/agents/registry.json

@@ -168,7 +168,7 @@
       "optionalProducesArtifacts": [],
       "ownership": "lead quality, false-positive patterns, keep/exclude rules, ability-to-pay checks, and production rubric translation only",
       "codex": {
-        "description": "Lead Fit Builder for campaign-backed lead filtering and rubric persistence after source approval.",
+        "description": "Lead Fit Builder for campaign-backed lead filtering and rubric persistence after review-batch import and filter approval.",
         "model": "gpt-5.5",
         "modelReasoningEffort": "high",
         "sandboxMode": "read-only",
@@ -179,7 +179,7 @@
         ]
       },
       "claude": {
-        "description": "Use proactively as Lead Fit Builder after lead source approval to persist campaign rubrics from campaign state.",
+        "description": "Use proactively as Lead Fit Builder after review-batch import and filter approval to persist campaign rubrics from campaign state.",
         "model": "inherit",
         "background": true,
         "maxTurns": 8,

package/agents/source-scout-linkedin-engagement.md

@@ -51,7 +51,8 @@ currentStep: "signal-discovery" })` before sampling so the watched Signal
    visible headline/display-name cues only. Do not enrich people during
    viability estimation.
 7. Compute capacity before recommending the source: source target good-fit
-   leads (default 300 unless the parent supplies a target), reachable engagers,
+   leads (default 150 for Signal Discovery unless the parent supplies a target),
+   reachable engagers,
    sampled ICP-fit rate as `n/N` plus an easy percentage/range, expected usable
    leads per 100 engagers before and after a conservative dedupe/cleanup
    factor, required engagers to scrape (`source target / fit rate`), average
@@ -93,8 +94,8 @@ Evidence standards:
   exist, how many sampled engagers looked in-ICP, how many good-fit prospects
   that implies per 100 engagers, how many usable prospects one right-content
   post should yield after cleanup, how many engagers must be scraped for the
-  300-good-fit source target, how many posts are needed for that source target,
+  150-good-fit source target, how many posts are needed for that source target,
   and which posts you would use. Also say the first campaign import remains the
-  bounded 15-row review batch.
+  bounded review batch.
 - If `fetch_post_engagers` is unavailable or fails, report that explicitly and mark the estimate lower-confidence.
 - Keep LinkedIn Engagement viable when selected posts can produce roughly 150+ ICP-fit warm prospects before final filtering, even if Sales Nav is more scalable.

package/agents/source-scout-prospeo-contact.md

@@ -27,7 +27,7 @@ Process:
 1. Read the campaign brief, source intake, kickoff doc, or lane prompt supplied by the parent.
 2. Identify whether this is domain/account targeting or broad persona expansion.
 3. For domain targeting, use or create the standalone `domainFilterId` before searching; never pass raw domains directly into `search_prospeo`.
-4. Run the narrowest useful Prospeo people preview and 1-2 refinements if quality or scale is unclear.
+4. Run the narrowest useful Prospeo people preview and 1-2 refinements if quality or scale is unclear. Check scale against a default 300+ good-fit target, capped at 2,500 source candidates unless the parent supplies a different target.
 5. Call out that Prospeo gives contact/account coverage but usually weaker LinkedIn intent than LinkedIn Engagement or Sales Nav activity slices.
 
 Return a concise structured result with:

package/agents/source-scout-sales-nav.md

@@ -28,8 +28,9 @@ Process:
 2. Preserve target role names with `CURRENT_TITLE` lookups; do not rely on seniority alone when the brief names concrete roles.
 3. When `lookup_sales_nav_filter` returns multiple title options, choose the closest semantic title match instead of the first result.
 4. Build a broad-but-reasonable baseline from role/title, geography, company size, industry/account context, and recent LinkedIn activity when relevant.
-5. Check scale against the source target good-fit lead count (default 300
-   unless the parent supplies a target). If raw preview volume or projected usable volume
+5. Check scale against the source target good-fit lead count (default 300+
+   for Sales Nav, capped at 2,500 source candidates unless the parent supplies a target).
+   If raw preview volume or projected usable volume
    is below target, do not present the tiny result as the scale fallback yet.
    Loosen nonessential filters in order: remove recent-activity first, widen
    adjacent title variants, widen geography/company-size constraints, and only

package/bin/sellable-install.mjs

@@ -40,7 +40,7 @@ function getMcpVersion() {
   } catch {}
   return "latest";
 }
-const CODEX_PLUGIN_VERSION = "0.1.32";
+const CODEX_PLUGIN_VERSION = "0.1.33";
 const CODEX_PLUGIN_COMPAT_VERSIONS = [
   "0.1.8",
   "0.1.9",
@@ -166,9 +166,7 @@ Auth:
 
 function printCreateCommandHint() {
   printBanner();
-  console.log(
-    `  ${C.bold}Run Sellable from Claude Code or Codex.${C.reset}`
-  );
+  console.log(`  ${C.bold}Run Sellable from Claude Code or Codex.${C.reset}`);
   console.log("");
   console.log(
     `  ${C.grey}The terminal command installs and verifies Sellable. Campaign creation runs inside your agent session so the workflow can use MCP tools, live state, and approval gates.${C.reset}`
@@ -182,12 +180,14 @@ function printCreateCommandHint() {
   printAgentBox("Using Claude Code?", "claude", [
     { label: "Campaign", command: "/sellable:create-campaign" },
     { label: "Identity", command: "/sellable:interview" },
+    { label: "Voice", command: "/sellable:load-voice" },
   ]);
   console.log("");
   console.log("");
   printAgentBox("Using Codex?", "codex", [
     { label: "Campaign", command: "$sellable:create-campaign" },
     { label: "Identity", command: "$sellable:interview" },
+    { label: "Voice", command: "$sellable:load-voice" },
   ]);
   console.log("");
   console.log(`   ${"─".repeat(63)}`);
@@ -630,8 +630,8 @@ clear business decisions, tradeoffs, and approval gates. Use product language:
 - "a couple setup choices", not \`request_user_input\`
 - "campaign brief", not prompt artifact
 - "lead source", not provider internals unless comparing source options
-- "I can create a draft shell for you to watch, but no leads import and nothing
-  sends until you approve", not mutation jargon
+- "I can create a draft shell for you to watch with approval gates before
+  sourcing", not mutation jargon
 
 When explaining lead-source decisions, show the concrete counts behind the
 logic: lanes searched, timeframe, raw result counts, finalist posts or preview
@@ -717,7 +717,7 @@ After every \`update_campaign({ campaignId, currentStep })\`, use
 \`get_campaign_navigation_state\` when available as a compact orientation check:
 match the saved campaign state to the expected watch-link step, explain the
 current state in one sentence, and only then continue. Sender selection belongs
-at Settings after message approval and 15-row validation. After message
+at Settings after message approval and review-batch validation. After message
 validation, use Settings to help the user connect or select a LinkedIn sender.
 Explain Slack reply review before launch. After sender selection, attach the
 recommended sequence and move the watched UI to Send. Do not start the campaign
@@ -822,7 +822,7 @@ the brief. This is only the client-prospect/bootstrap identity for
 Do not call \`mcp__sellable__list_senders\`, \`mcp__sellable__get_sender\`, or
 surface connected/missing sender state during setup, brief, source, filter, or
 message review. Sender availability belongs only to the Settings/final launch
-handoff after message approval and the 10-lead validation sample.
+handoff after message approval and the review-batch validation sample.
 
 If the invocation or user answer includes an existing \`clientProspectId\`, keep
 it as the preferred \`create_campaign\` identity input. If it includes a LinkedIn
@@ -959,9 +959,9 @@ updates.
    until \`hasMore=false\`. The create-campaign message-review safety gate is a
    supplemental approval checklist, not a replacement for the long prompt. Use
    campaign state, campaign brief content, selected source state, and imported
-   review-batch rows as the source of truth; do not read stale local markdown,
-   inspect the database directly, or synthesize local validation artifacts from
-   general knowledge.
+   review-batch rows as the source of truth; do not read stale local markdown
+   such as \`message-validation.md\`, inspect the database directly, or synthesize
+   local validation artifacts from general knowledge.
 5. Create the campaign shell early with the v1 brief so the user can open the
    watch link and see useful setup state immediately. Import only the first
    bounded review batch after the source is attached to the campaign; do not
@@ -979,7 +979,7 @@ updates.
    path before that template/token approval. AI Generated uses the
    product's AI-generated path and cancels or ignores the background template
    draft.
-6. The main thread owns watch navigation: use \`filter-choice\` after the 15-row
+6. The main thread owns watch navigation: use \`filter-choice\` after the review
    review batch, \`create-icp-rubric\` as soon as filters are approved,
    \`apply-icp-rubric\` after rubrics save, \`messages\` for the Use Template / AI Generated mode choice,
    \`auto-execute-messaging\` for approved message work or the product's
@@ -1089,6 +1089,55 @@ then retry \`get_subskill_prompt\`.
 `;
 }
 
+function loadVoiceSkillMd() {
+  return `---
+name: load-voice
+description: Load Sellable voice/company memory for answering questions, writing posts, applications, replies, and reviews in the user's voice.
+allowed-tools:
+  - mcp__sellable__get_subskill_prompt
+  - mcp__sellable__search_subskill_prompts
+  - Read
+  - Glob
+  - Grep
+---
+
+# Sellable Load Voice
+
+Use this as the customer-facing entrypoint for loading Sellable voice/company
+memory before drafting, answering, rewriting, reviewing, or calibrating copy in
+the user's voice.
+
+## Bootstrap
+
+MCP prompt access and local read tools are required. Do not run shell commands,
+use \`npm\`, \`node\`, local harness scripts, or read local prompt files to
+emulate this workflow.
+
+If the Sellable MCP prompt tools are unavailable, stop and say this is a Codex
+install/reload problem. Tell the user to run
+\`npx -y ${INSTALL_PACKAGE_SPEC} --host all\`, fully quit and reopen Codex
+Desktop, then start a new thread.
+
+## Execute Workflow
+
+1. Load the canonical prompt via
+   \`mcp__sellable__get_subskill_prompt({ subskillName: "load-voice" })\`.
+   If the response has \`hasMore=true\`, continue with \`nextOffset\` until
+   \`hasMore=false\`.
+2. Follow the canonical prompt exactly. Read the relevant
+   \`.sellable/configs/**\` memory files it names.
+3. Apply the loaded memory silently to the user's requested writing or answer.
+   If the user only asked to load voice, summarize the active rules briefly and
+   ask what they want drafted, answered, or reviewed.
+
+## MCP Prompt Fallback
+
+If exact subskill lookup fails, use
+\`mcp__sellable__search_subskill_prompts({ query: "load-voice", includePublic: true })\`,
+then retry \`get_subskill_prompt\`.
+`;
+}
+
 function createCampaignSoulMd() {
   return `# Sellable Campaign GTM Engineer Soul
 
@@ -1283,6 +1332,12 @@ function codexPluginSkills() {
       description: "Build durable identity and company memory",
       skillMd: interviewSkillMd(),
     },
+    {
+      dir: "sellable-load-voice",
+      displayName: "Sellable Load Voice",
+      description: "Load voice memory for writing and answers",
+      skillMd: loadVoiceSkillMd(),
+    },
   ];
 }
 
@@ -2129,6 +2184,7 @@ function printNextSteps(installedHosts, authReused) {
     printAgentBox("Using Claude Code?", "claude", [
       { label: "Campaign", command: "/sellable:create-campaign" },
       { label: "Identity", command: "/sellable:interview" },
+      { label: "Voice", command: "/sellable:load-voice" },
     ]);
     console.log("");
     console.log("");
@@ -2137,6 +2193,7 @@ function printNextSteps(installedHosts, authReused) {
     printAgentBox("Using Codex?", "codex", [
       { label: "Campaign", command: "$sellable:create-campaign" },
       { label: "Identity", command: "$sellable:interview" },
+      { label: "Voice", command: "$sellable:load-voice" },
     ]);
     console.log("");
   }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@sellable/install",
-  "version": "0.1.104",
+  "version": "0.1.106",
   "type": "module",
   "description": "One-command installer for Sellable MCP in Claude Code and Codex",
   "bin": {

package/skill-templates/create-campaign.md

@@ -86,7 +86,7 @@ or tool discovery. Start in product language:
 ```text
 I’ll help you launch this as a Sellable campaign. First I’ll resolve the
 client/company this campaign is for, then I’ll turn that into a campaign brief
-before any leads are imported or anything can send.
+before we move into lead sourcing.
 ```
 
 If a linked/local skill file is stale or missing, silently use the installed
@@ -102,8 +102,14 @@ clear business decisions, tradeoffs, and approval gates. Use product language:
 - "a couple setup choices", not `request_user_input`
 - "campaign brief", not prompt artifact
 - "lead source", not provider internals unless comparing source options
-- "I can create a draft shell for you to watch, but no leads import and nothing
-  sends until you approve", not mutation jargon
+- "I can create a draft shell for you to watch with approval gates before
+  sourcing", not mutation jargon
+
+Approval and safety copy should be tasteful. State what the current approval
+covers once, in one short sentence, then move on. Do not append repeated
+"nothing starts / no leads import / no sending" disclaimers to routine progress
+updates. Use positive gate language like "Next gate: selected-post scrape" or
+"Approval covers scouting/search only" instead of long negative lists.
 
 When explaining lead-source decisions, show the concrete counts behind the
 logic: lanes searched, timeframe, raw result counts, finalist posts or preview
@@ -127,7 +133,7 @@ authorizes scouting/search only. The gate should say:
 - why that lane fits the buyer, offer, and likely public activity
 - what will be tested next
 - the fallback lane if relevant posts or ICP engagement look thin
-- that approval authorizes scouting/search only, not lead import or sending
+- what approval covers in one concise line
 
 Do not surface blanket source heuristics as product copy. Make the
 recommendation specific to the campaign. If Signal Discovery is recommended,
@@ -142,6 +148,21 @@ Signal Discovery, name how many selected posts will be scraped, the target
 engager/source-candidate volume, and the bounded review-batch size. For Sales
 Nav or Prospeo, name the specific approved import lane. Do not call
 `import_leads` or `confirm_lead_list` until this second approval is granted.
+For Signal Discovery, the customer-facing approval card must use the exact
+action shape "Approve scraping N Signal Discovery posts?" and the chat summary
+should be a compact `## Source Recommendation` block with:
+
+- good-fit target: about 150 prospects after cleanup, enrichment, and filters
+- source-candidate plan: about 1,000 raw engagers using a conservative 15%
+  fit-rate assumption unless sampled data supports a different number
+- review checkpoint: import the first 25 leads for fit and message review
+- a selected-post table with post author/topic, why it fits, and visible
+  engagement
+- total visible pool and estimated good-fit pool
+- first pass: build the source list, then import only the review
+  batch
+- fallback: switch to Sales Nav recent activity if the review batch is
+  vendor-heavy, agency-heavy, or off-ICP
 
 When the user has not supplied a source and multiple source angles are viable,
 scout those angles as independent branches when the host can actually do it:
@@ -168,19 +189,17 @@ which posts are being sampled in the watched app. The watch guide should say
 that we are pulling sample engagers from these posts to confirm the ICP is
 actually engaging and the source is viable.
 
-After find-leads returns a lead source and the user approves it, use the same
-registry pattern for the two post-lead branches. The create-campaign-v2 subskill
-calls `get_post_find_leads_scout_registry`, then launches the returned
-filter-leads scout and message-generation scout together when real subagents are
-available and the current session exposes the returned names. Message
-generation is the provisional Message Draft Builder: it may start after the
-review batch exists, including while the user is on filter choice, but workflow
-cell execution must wait. After rubrics are saved, do not queue enrichment,
-filtering, or Generate Message cells until the user approves the message
-template. AI Generated is an explicit opt-out that cancels or ignores the
-background template draft. If the post-lead agents are absent, the main thread
-still orchestrates the same branches from the compact context with MCP
-tools/assets.
+After the bounded review batch exists, use the same registry pattern for
+post-lead work. The create-campaign-v2 subskill calls
+`get_post_find_leads_scout_registry`, then launches the returned
+message-generation scout immediately when real subagents are available and the
+current session exposes the returned name. If the user chooses filters, launch
+the filter-leads scout then and join it with the already-running or completed
+message draft before message review. Workflow cell execution still waits for
+filter and template approval. AI Generated is an explicit opt-out that cancels
+or ignores the background template draft. If the post-lead agents are absent,
+the main thread still orchestrates the same branches from the compact context
+with MCP tools/assets.
 
 Use rendered Markdown for user review surfaces, not fenced code blocks. Keep
 lines short, use indexed section labels and bullets, and translate internal
@@ -233,7 +252,7 @@ After every `update_campaign({ campaignId, currentStep })`, use
 `get_campaign_navigation_state` when available as a compact orientation check:
 match the saved campaign state to the expected watch-link step, explain the
 current state in one sentence, and only then continue. Sender selection belongs
-at Settings after message approval and 15-row validation. After message
+at Settings after message approval and review-batch validation. After message
 validation, use Settings to help the user connect or select a LinkedIn sender.
 Explain Slack reply review before launch. After sender selection, attach the
 recommended sequence and move the watched UI to Send. Do not start the campaign
@@ -412,7 +431,7 @@ First I’ll resolve the client/company this campaign is for. I’ll use that
 context to choose the target, offer, proof, and lead source.
 
 Then I’ll turn that into a campaign brief for you to approve before any leads
-are imported or anything can send.
+are sourced.
 ```
 
 Do not silently ask Codex intake or approval questions as plain chat when
@@ -596,9 +615,9 @@ updates.
    until `hasMore=false`. The create-campaign message-review safety gate is a
    supplemental approval checklist, not a replacement for the long prompt. Use
    campaign state, campaign brief content, selected source state, and imported
-   review-batch rows as the source of truth; do not read stale local markdown,
-   inspect the database directly, or synthesize local validation artifacts from
-   general knowledge.
+   review-batch rows as the source of truth; do not read stale local markdown
+   such as `message-validation.md`, inspect the database directly, or synthesize
+   local validation artifacts from general knowledge.
 5. Create the campaign shell early with the v1 brief so the user can open the
    watch link and see useful setup state immediately. Import only the first
    bounded review batch after the source is attached to the campaign; do not
@@ -609,9 +628,11 @@ updates.
    call `mcp__sellable__update_campaign({ campaignId, enableICPFilters: true, currentStep: "create-icp-rubric", watchNarration })`
    so the watched app moves to Filter Rules while rubrics are drafted/saved.
    After rubrics save, move the watched app to `apply-icp-rubric` / Filter
-   Leads and say the fit rules are saved; approve the message template next;
-   after approval, queue the bounded review-batch
-   `enrichCellId` cells to kick off enrichment/filtering.
+   Leads and say the fit rules are saved; approve the message template next
+   while the browser stays on Filter Leads. After approval, save the template
+   to the campaign brief, then queue the bounded review-batch `enrichCellId`
+   cells to kick off enrichment/filtering. Move to Messages only after at
+   least one review row passes and Generate Message cells are running or ready.
    Product Generate Message cells must not run from the background template
    path before that template/token approval.
    Do not ask the user to approve the brief before shell creation unless they
@@ -620,15 +641,15 @@ updates.
    `mcp__sellable__update_campaign({ campaignId, currentStep })` before major
    visible work so the user can watch progress in the app: `create-offer` for
    the brief, `pick-provider` or the selected provider step while sourcing,
-   `filter-choice` after the 15-row review batch, `create-icp-rubric` as soon
-   as filters are approved, `apply-icp-rubric` after rubrics save, `messages`
-   for the Use Template / AI Generated mode
-   choice, `auto-execute-messaging` for approved
-   message work or the product's AI-generated path, `awaiting-user-greenlight`
-   for the final handoff, `settings` for sender selection, `sequence` after
-   sender attach, and `send` once the recommended sequence is attached.
-   `validate-sample` is recovery/legacy only if reached. Do not advance the
-   step backward.
+   `filter-choice` after the review batch, `create-icp-rubric` as soon
+   as filters are approved, `apply-icp-rubric` after rubrics save and while
+   waiting for message-template approval, `validate-sample` while the approved
+   template unlocks bounded enrichment/filter scoring on Filter Leads,
+   `auto-execute-messaging` after at least one row passes and review-batch
+   messages are being generated or reviewed, `awaiting-user-greenlight` only
+   after generated review-batch messages are approved, `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
    campaign table. Do not use disk files as the post-mint source of truth.
 8. Do not ask the user to run another command.