@sellable/install 0.1.231 → 0.1.233

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -37,7 +37,7 @@ function getInstallVersion() {
37
37
  }
38
38
  }
39
39
 
40
- const CODEX_PLUGIN_VERSION = "0.1.48";
40
+ const CODEX_PLUGIN_VERSION = "0.1.50";
41
41
  const CODEX_PLUGIN_COMPAT_VERSIONS = [
42
42
  "0.1.8",
43
43
  "0.1.9",
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@sellable/install",
3
- "version": "0.1.231",
3
+ "version": "0.1.233",
4
4
  "type": "module",
5
5
  "description": "One-command installer for Sellable MCP in Claude Code and Codex",
6
6
  "bin": {
@@ -106,12 +106,19 @@ Short form: nested filesystem prompt fallback is a failed UAT.
106
106
  Package-backed tool surface: use `mcp__sellable` only for workspace selection,
107
107
  campaign/table/source/message mutations, and product-visible readbacks. Do not
108
108
  use `mcp__sellable_admin`, admin workspace setters, direct DB tools, Prisma, or
109
- SQL as part of command execution proof. Before any worker mutation, record
109
+ SQL as part of command execution proof. Do not use built-in web search,
110
+ browser search, `web.run`, or any external browsing/search tool for sender or
111
+ company research in this command; use Sellable MCP research tools such as
112
+ `get_sender`, `fetch_linkedin_profile`, `fetch_linkedin_posts`,
113
+ `fetch_company`, `fetch_company_posts`, and `complete_sender_research`. If those
114
+ MCP tools do not provide enough proof, write a blocked receipt with the missing
115
+ proof instead of browsing externally. Before any worker mutation, record
110
116
  normal-surface workspace proof with
111
117
  `set_active_workspace({ workspaceId, userConfirmed:true })` or the equivalent
112
118
  `mcp__sellable` call, then reread the active workspace.
113
119
  Short form: mcp__sellable only.
114
120
  Short form: Do not use `mcp__sellable_admin`.
121
+ Short form: Do not use web search or browser search.
115
122
 
116
123
  Plan shape:
117
124
 
@@ -342,7 +349,12 @@ not a parent-thread summary. The receipt must include:
342
349
  `firstMessageTemplate`; if it uses `deliveryFormatIncluded`,
343
350
  `tokenRulesIncluded`, `hardAvoidsIncluded`, or `sourceUseRulesIncluded`, copy
344
351
  those to canonical `deliveryFormat`, `tokenRules`, `hardAvoids`, and
345
- `sourceUseRules`.
352
+ `sourceUseRules`. Before the worker exits, the durable JSON must contain the
353
+ exact canonical leaf keys `briefHash`, `deliveryFormat`, `tokenRules`,
354
+ `hardAvoids`, `sourceUseRules`, and `firstMessageTemplate` inside
355
+ `createCampaignStepReceipt.campaignBriefReceipt`; do not leave those only in
356
+ narrative text, `briefIncludes`, `currentApprovedBriefHash`, or other aliases
357
+ for the parent/verifier to infer.
346
358
  - `sourceDecisionReceipt`: selected source/list id, provider, cursor/inventory
347
359
  state, import/copy caps, explicit `dedupeDncProviderExclusions:true` or an
348
360
  equivalent exclusions object, row ids or row-count evidence (`rowCount`,
@@ -350,7 +362,12 @@ not a parent-thread summary. The receipt must include:
350
362
  before the current source was proven exhausted or insufficient. If the raw
351
363
  provider output uses `sourceProvider`, `sourceListId`, `reviewBatchRowIds`, or
352
364
  `rationale`, copy them to canonical `provider`, `sourceId`/`sourceListId`,
353
- `rowIdsHash`/`rowCount`, and `sourceLadderReason`.
365
+ `rowIdsHash`/`rowCount`, and `sourceLadderReason`. Before the worker exits,
366
+ the durable JSON must contain exact canonical source proof fields:
367
+ `dedupeDncProviderExclusions`, `sourceLadderReason`, and at least one row
368
+ evidence field such as `rowCount`, `rowIdsHash`, or `rowEvidence`; do not
369
+ leave those only in `searchSignalsProof`, `importProof`, `coldFallbackProof`,
370
+ or narrative text for the parent/verifier to infer.
354
371
  - `filterDecisionReceipt`: saved/applied filter ids or explicit skip-filter
355
372
  decision with current campaign/table basis.
356
373
  - `messageDraftingReceipt`: the Message Drafting proof described below,
@@ -400,6 +417,16 @@ before writing JSON; do not leave only raw objects like
400
417
  `get_campaign_message_preparation_status`, `attachRecommendedSequenceResult`,
401
418
  `finalCampaignRead`, or `finalTableRead`.
402
419
  Self-check shorthand: top-level `planRevision`, `actionId`, `laneKey`, `laneType`, `workspaceId`, and `senderIds` must exist before worker completion.
420
+ Self-check shorthand: `campaignBriefReceipt.briefHash`,
421
+ `campaignBriefReceipt.deliveryFormat`, `campaignBriefReceipt.tokenRules`,
422
+ `campaignBriefReceipt.hardAvoids`, `campaignBriefReceipt.sourceUseRules`, and
423
+ `campaignBriefReceipt.firstMessageTemplate` must exist as exact canonical keys
424
+ before worker completion.
425
+ Self-check shorthand: `sourceDecisionReceipt.dedupeDncProviderExclusions`,
426
+ `sourceDecisionReceipt.sourceLadderReason`, and one of
427
+ `sourceDecisionReceipt.rowCount`, `sourceDecisionReceipt.rowIdsHash`, or
428
+ `sourceDecisionReceipt.rowEvidence` must exist as exact canonical keys before
429
+ worker completion.
403
430
  Self-check shorthand: `messageDraftingReceipt.promptLoadedToHasMoreFalse`,
404
431
  `requiredAssetsLoaded`, `validationLoaded`, `reviewBatchRowHash`, and
405
432
  `messageDraftRecommendation` must exist before worker completion.