@sellable/mcp 0.1.271 → 0.1.273

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@sellable/mcp",
-  "version": "0.1.271",
+  "version": "0.1.273",
   "type": "module",
   "description": "Sellable MCP server for Claude Code and Codex campaign workflows",
   "main": "dist/index.js",

package/skills/create-post/SKILL.md

@@ -336,6 +336,13 @@ write from the user's idea and context. The worker may write in a voice-inspired
 style for the private variant, but it must not claim the creator authored it,
 copy source wording, copy personal proof, or invent facts about the user.
 
+Voice variants must model **writing mechanics**, not just a broad persona.
+Before drafting, each worker must infer what "in this creator's voice" concretely
+means for this idea: target word-count band, line and paragraph rhythm, hook
+posture, recurring phrase families, vocabulary level, proof pattern, formatting
+habits, close style, and normal amount of directness or humor. If every variant
+comes back as the same long LinkedIn essay with different labels, the lab failed.
+
 The final orchestrator does not publish a collage of creator personas. It takes
 the best hooks, pressure, structure, rhythm, and proof moves from the role-played
 variants, then synthesizes the most compelling post for Christian/Sellable.
@@ -388,22 +395,26 @@ Default workflow:
    repeated patterns, engagement quality, and follower-adjusted signal when
    follower counts are available. Do not return the full scoring table.
 6. Each worker must build a `voice_model` before drafting:
-   - representative length range for similar posts
+   - target word-count band and normal variance for similar posts
    - average paragraph/line length and spacing rhythm
    - hook posture and opening moves
-   - common vocabulary, recurring phrases, and taboo phrases
+   - common vocabulary, recurring phrase families, and taboo phrases
    - proof style: story, numbers, teardown, checklist, confession, challenge
    - formatting: bullets, numbering, questions, equations, parentheticals
    - close style: command, reflection, CTA, punchline, open loop
    - factual boundaries: what must come from Christian/Sellable context
 7. Each worker writes the complete post that this creator would likely write
    from the user's idea and context. Match the inferred length/rhythm/style; do
-   not compress to a generic short draft.
+   not compress to a generic short draft, and do not stretch a short-style
+   creator into a generic long post.
 8. Each worker returns a compact voice variant:
    - person and profile URL
    - worker status
    - recent post count sampled
    - voice model, max 8 bullets
+   - mechanics receipt: target length, actual length, line rhythm,
+     vocabulary markers used, formatting markers used, close style, and
+     `mechanics_match: pass | weak | fail`
    - complete post variant in that creator's inferred voice and representative
      length
    - hook options, max 3
@@ -427,6 +438,10 @@ Hard rules:
   by the creator.
 - Do not copy outside wording, proof, jokes, personal stories, or status.
 - Do not invent user proof just because it would fit the creator's style.
+- Do not treat "voice" as a vibe label. Voice means concrete mechanics: length,
+  line rhythm, vocabulary, hook posture, proof style, formatting, and close.
+- Do not publish or save a `ready` draft when voice variants are long generic
+  essays with swapped creator names instead of person-specific mechanics.
 - Do not let a large-audience creator win only because of reach. Use engagement
   per 1k followers when follower counts are available, and mark confidence when
   follower data is missing.
@@ -591,6 +606,7 @@ Visible Flow Trace
 - worker status per person:
 - recent posts sampled:
 - voice model:
+- mechanics receipts:
 - complete voice variants:
 - follower-adjusted signal when available:
 - hook/body moves worth stealing:
@@ -608,6 +624,11 @@ Visible Flow Trace
 - proof gate:
 - voice gate:
 - anti-AI gate:
+- local finalizer pass:
+- turn separators added:
+- contractions applied:
+- lowercase pass:
+- abstract-to-concrete rewrites:
 - mobile preview gate:
 - template adaptation gate:
 - hook-to-body repayment:
@@ -1118,8 +1139,11 @@ Each worker must return:
 - recent post count sampled
 - follower count when available
 - normalization confidence
-- voice model, max 8 bullets, including representative length, rhythm,
-  vocabulary, formatting, proof style, and close style
+- voice model, max 8 bullets, including target length, rhythm, vocabulary,
+  formatting, proof style, and close style
+- mechanics receipt, including target word-count band, actual word count,
+  paragraph/line rhythm, vocabulary markers, formatting markers, close style,
+  and `mechanics_match: pass | weak | fail`
 - complete role-played post variant in the creator's inferred voice and
   representative length
 - hook options, max 3
@@ -1138,6 +1162,7 @@ Variant labels must be:
 ```text
 voice_variant_from: <person>
 role_play_basis: public posts sampled + configured Reason lane
+mechanics_receipt: target length | actual length | rhythm | vocabulary markers | formatting markers | close | mechanics_match
 post_variant: <complete variant>
 steal_for_final: hook | structure | proof order | rhythm | close move
 drop_for_final: copied wording | borrowed personal proof | fake user fact
@@ -1156,6 +1181,7 @@ Before drafting, record:
 - `actualVoiceVariantCount`
 - `missingVoiceVariants`
 - `weakVoiceVariants`
+- `mechanicsReceipts`
 - `selectedSynthesisIngredients`
 - `voiceVariantsGeneratedBeforeFinalProse: yes | no`
 - `workerExecutionMode: background_agents | sequential_fallback`
@@ -1175,6 +1201,7 @@ actualVoiceVariantCount: <number>
 
 1. <person>
 - worker_status:
+- mechanics_match:
 - draft_review:
 - strongest_move:
 - weakness_or_drop:
@@ -1221,6 +1248,42 @@ either completed for every active configured influencer or the validation
 receipt has an explicit user opt-out reason. A light source-research summary
 does not satisfy this gate.
 
+## Step 3.5: Local Finalizer Pass
+
+Run this pass after the canonical draft exists and before validation and save.
+
+1. Add `--` between each major turn in the post so the story is easier to read.
+   Use it to separate turns such as hook, context, problem, mechanism, product
+   proof, concrete example, and CTA. Do not add `--` inside a tight list where
+   it would make the post harder to scan.
+2. Use contractions in the draft body by default: `don't`, `isn't`, `that's`,
+   `it's`, `can't`, `we're`, `you're`, and similar natural phrasing. Keep the
+   non-contracted form only when the user explicitly wrote it that way, when a
+   quoted line must stay exact, or when the contraction would sound unnatural.
+3. Keep optional headings, section labels, and line starts lowercase by default
+   when the post still reads naturally. Preserve proper nouns, acronyms, product
+   names, and required capitalization such as `Claude`, `LinkedIn`, `Codex`,
+   `Sellable.dev`, `Clay`, `HeyReach`, and `ICP`.
+4. Run a Harry Dry-style concrete-language audit:
+   - highlight every abstract or vague word/phrase in the validation receipt
+     before saving
+   - replace each one in the draft with a concrete object, action, number,
+     example, screen, list, person, place, or visible workflow step
+   - prefer words a reader can picture, such as `lead list`, `send settings`,
+     `reply inbox`, `follow-up`, `team post`, `warm LinkedIn prospect`,
+     `reply handling`, or `one sentence in Claude Code`
+   - do not make a phrase concrete by inventing proof, metrics, customers, or
+     product behavior the user did not provide
+5. Record this pass in the validation receipt as `localFinalizerPass` with:
+   - `turnSeparatorsAdded: yes | no` and a short reason
+   - `contractionsApplied: yes | no`
+   - `lowercasePass: yes | no` with preserved proper nouns listed
+   - `abstractToConcreteRewrites`, listing each abstract phrase, why it was
+     weak, and the concrete replacement used
+
+If this pass cannot be completed, save the draft as `needs_revision` instead of
+`ready`, and explain which finalizer requirement failed.
+
 ## Step 4: Validation
 
 Use `references/post-validation.md`.
@@ -1238,7 +1301,8 @@ Every saved draft needs a validation receipt with:
 - pre-draft narrative outline
 - selected source template and no-copy adaptation rationale
 - thought leader voice variant lab, including skipped/opt-out reason when
-  omitted and expected-vs-actual variant counts for the active configured list
+  omitted, expected-vs-actual variant counts for the active configured list,
+  and per-person mechanics receipts
 - post positioning breakdown
 - viral-post outline
 - hook-to-body promise map
@@ -1258,6 +1322,8 @@ Every saved draft needs a validation receipt with:
 - outbound AI-tell audit findings
 - hook preview pass/warn/fail status and compact fallback when warned
 - finalizer changes
+- local finalizer pass with turn separators, contractions, lowercase pass, and
+  abstract-to-concrete rewrites
 - blocked/retry-needed reasons, if any
 
 If validation fails, do not save the draft as `ready`. Save as `needs_revision` or return blocked/retry-needed.

package/skills/create-post/references/hook-research-playbook.md

@@ -718,6 +718,13 @@ missing variants, weak variants, and selected synthesis ingredients. Ready
 drafts require one voice variant per active configured person unless the user
 explicitly opted out or supplied a smaller named subset.
 
+Voice variants must model writing mechanics, not just a broad creator lens.
+Each worker has to define what "in this person's voice" means for this idea:
+target word-count band, line and paragraph rhythm, hook posture, recurring
+phrase families, vocabulary level, proof pattern, formatting habits, close
+style, and normal directness or humor. If multiple creator variants all become
+the same long LinkedIn essay with different labels, mark the lab weak or failed.
+
 Use one bounded background worker per person when the host supports background
 agents. This is required for Codex/Claude hosts that expose background agents or
 Task workers. If the host does not expose a worker mechanism, process the list
@@ -755,10 +762,10 @@ Worker steps:
    - adaptation fit for Christian/Sellable voice
    - lead-magnet, giveaway, and status-only penalties
 5. Build a `voice_model` before drafting. It must infer:
-   - representative post length range for similar ideas
+   - target word-count band and normal variance for similar ideas
    - paragraph and line rhythm
    - hook posture and opening moves
-   - vocabulary and recurring phrases
+   - vocabulary and recurring phrase families
    - formatting habits: bullets, numbering, equations, parentheticals, spacing
    - proof style: story, metric, teardown, confession, field guide, challenge
    - close style: command, reflection, CTA, punchline, or open loop
@@ -768,7 +775,8 @@ Worker steps:
    hook tension, pacing, section shape, directness, humor level, proof style,
    formatting, and editorial instincts.
 7. Write the complete post that this creator would likely write from the user's
-   selected premise and context. Do not compress it into a short generic draft.
+   selected premise and context. Do not compress it into a short generic draft
+   or stretch a short-style creator into a generic long post.
 8. Identify what the final synthesis should steal and what it must drop.
 9. Keep factual claims grounded in the user's context. Do not invent user proof.
 
@@ -796,7 +804,15 @@ worker_status: pass | weak | retry_needed | blocked
 recent_posts_sampled:
 follower_count: <number | unavailable>
 normalization_confidence: high | medium | low
-voice_model: [max 8 bullets covering length, rhythm, vocabulary, formatting, proof style, close]
+voice_model: [max 8 bullets covering target length, rhythm, vocabulary, formatting, proof style, close]
+mechanics_receipt:
+- target_word_count_band:
+- actual_word_count:
+- paragraph_line_rhythm:
+- vocabulary_markers:
+- formatting_markers:
+- close_style:
+- mechanics_match: pass | weak | fail
 hook_options: [max 3]
 post_variant: <complete post in the creator's inferred voice and representative length>
 steal_for_final: [max 5 bullets]

package/skills/create-post/references/post-file-contract.md

@@ -118,6 +118,10 @@ Draft files must preserve:
   promise, concrete examples, abstractions to remove, and draft risks
 - validation receipt, including LinkedIn preview pass/warn/fail status and
   compact fallback when the selected hook carries a warning
+- local finalizer pass: whether `--` turn separators were added, whether
+  contractions were applied, whether lowercase style was applied where safe,
+  preserved proper nouns/acronyms, and abstract-to-concrete rewrites with the
+  before/after phrases used in the final draft
 - visible flow trace when the user asked for whole-flow, debug, or step-by-step
   mode, including checkpoint statuses, quality break, downstream effect, and
   whether the trace was shown before saving

package/skills/create-post/references/post-validation.md

@@ -125,6 +125,41 @@ After the first draft:
 9. preserve the user's actual story and point
 10. remove AI tells
 11. re-check LinkedIn preview fit after edits
+12. add `--` between major story turns when it improves scanability
+13. use natural contractions by default
+14. keep optional headings, section labels, and line starts lowercase when safe
+15. run the concrete-language audit below and replace abstract phrasing before
+    saving
+
+### Local Finalizer Pass
+
+Before a draft can be saved as `ready`, record a local finalizer pass. This is
+the last copy pass after the draft body exists and before `save_post_draft`.
+
+Record:
+
+- `turnSeparatorsAdded`: yes | no
+- `turnSeparatorReason`: where `--` was inserted or why it was skipped
+- `contractionsApplied`: yes | no
+- `contractionsChanged`: examples such as `do not -> don't`, `is not -> isn't`,
+  `that is -> that's`, or `it is -> it's`
+- `lowercasePass`: yes | no
+- `preservedCapitalization`: proper nouns, acronyms, product names, and required
+  capitalization such as `Claude`, `LinkedIn`, `Codex`, `Sellable.dev`, `Clay`,
+  `HeyReach`, and `ICP`
+- `abstractToConcreteRewrites`: every abstract phrase flagged, why it was weak,
+  and the concrete replacement used
+
+Save as `needs_revision` when:
+
+- major story turns are visually hard to follow and no `--` separators or other
+  clear turn markers were added
+- the draft keeps stiff non-contracted phrasing without a user/source reason
+- optional headings or labels stay title-cased in a way that clashes with the
+  user's lowercase post style
+- meaningful abstract phrases remain in the draft after the concrete-language
+  audit
+- the validation receipt does not show what changed during this finalizer pass
 
 ## Visible Flow Trace Audit
 
@@ -260,7 +295,9 @@ or practical next action.
 ## Abstraction-To-Concrete Rewrite Audit
 
 Record every meaningful abstraction found in the draft and the concrete
-replacement. Do not merely say "made concrete."
+replacement. Do not merely say "made concrete." Use a Harry Dry-style standard:
+if the reader cannot picture the word on a screen, in a list, in a sentence they
+would type, or as a workflow step someone could perform, make it more concrete.
 
 Use this format:
 
@@ -279,11 +316,25 @@ Common abstractions to flag:
 - vague nouns such as `signal`, `market`, `attention`, `system`, `quality`,
   `context`, `machine`, or `workflow` when the sentence does not name the actual
   source, action, number, or object
+- vague product words such as `orchestrator`, `campaign object`, `GTM
+  superpowers`, `agentic`, `AI-native`, or `platform` when the sentence does not
+  name the visible work being done
 - broad persona labels used as lead sources
 - "warm" without explaining what made the source warm
 - "offer" without explaining why the recipient would believe or want it
 - "learning" without naming what metric changes the next decision
 
+Concrete replacements should name visible things, for example:
+
+- `lead list`
+- `send settings`
+- `reply inbox`
+- `follow-up`
+- `team post`
+- `warm LinkedIn prospect`
+- `reply handling`
+- `one sentence in Claude Code`
+
 Do not make language concrete by inventing facts. If the concrete replacement
 requires missing proof, ask the user or save as `needs_revision`.
 
@@ -356,8 +407,11 @@ Record:
 - `workerExecutionMode`: `background_agents` | `sequential_fallback`
 - `backgroundWorkerIds`: worker IDs or labels when available
 - `thoughtLeaderList`: names and profile URLs or handles
-- `voiceModels`: inferred length, rhythm, vocabulary, formatting, proof style,
-  and close style per person
+- `voiceModels`: inferred target length, rhythm, vocabulary, formatting, proof
+  style, and close style per person
+- `mechanicsReceipts`: target word-count band, actual word count, line rhythm,
+  vocabulary markers, formatting markers, close style, and
+  `mechanics_match: pass | weak | fail` per person
 - `voiceVariants`: one representative-length role-played post per person, or
   `not_available` with reason
 - `normalizationConfidence`: high | medium | low
@@ -380,6 +434,12 @@ when variants are replaced by a generic influencer summary, or when voice
 variants were generated after final prose instead of before final prose. Return
 `retry-needed` when a tool/package/search failure prevents fetching configured
 profiles or posts.
+
+Save as `needs_revision` when the variants do not demonstrate person-specific
+writing mechanics, when `mechanicsReceipts` are missing, or when multiple
+variants are effectively the same long LinkedIn essay with different creator
+labels.
+
 Only mark `thought_leader_adaptation: not_used` when the user explicitly opted
 out before drafting.