@elitedcs/ghl-mcp 3.85.0 → 3.87.0

package/CHANGELOG.md

@@ -1,5 +1,19 @@
 # Changelog
 
+## 3.87.0 — report cards: every finished secret-shopper check is graded A to F
+
+A finished check now ends in a report card: a letter grade and a score out of 100, worked out the same way on every report. Every point lost is listed with its reason, and a serious failure caps the grade (a step that silently does nothing can't score above a C). The report says exactly what it checked, and a check that could not see everything is shown grey and marked "Incomplete check", never in grade colour. One root cause costs points once: messages stuck behind a failed step or a draft workflow are listed, not charged again.
+
+The report card is its own page, made to save as a PDF or hand to a client: the verdict and grade, what to fix with the change each fix makes, the grade after the fixes (projected until the check runs again), how the score was reached, every step and what the lead received, and what was not tested. It prints cleanly on US Letter with no card or paragraph split across pages. In chat, the summary now starts with the grade.
+
+## 3.86.0 — fast-forward: test a two-week nurture in minutes
+
+Tick "Fast-forward" when you start the secret shopper and it makes a temporary, clearly named copy of each workflow you chose, with every step kept and every wait cut to 1 minute, runs the test lead through the copy, and deletes the copy afterwards. A nurture that takes two weeks in real life is checked in about ten minutes, step by step, with every finding named after your original workflow and step. The original is only ever read, never changed, and the report lists how long each wait really is and any quiet hours the copy skipped.
+
+Nothing reaches your staff during the test. Staff alerts (in-app or by email), tasks, and lead or card assignments go to one person you choose, marked TEST. The start screen reads the chosen workflows and lists every outside service they send the lead to (a CRM, Zapier, a sheet) as a tick box with a plain explanation; the test lead is only sent to the ones you tick. For now fast-forward handles straight-line workflows with time waits; a workflow with if/else branches, hand-offs to other workflows, other kinds of waits, text-message staff alerts, or a webhook address built from a merge field is refused with the reason, before anything is created. Temporary copies left by an interrupted check are cleaned up automatically, and only ever the ones this computer recorded making.
+
+Also fixed in the secret shopper: GoHighLevel sometimes lets a lead out of a wait several minutes late. The check now waits up to 10 extra minutes for a wait that should already be over, and anything still held behind it is reported as "held by GoHighLevel, run it again", never as "never sent", and the headline never reads all-clear while messages are held.
+
 ## 3.85.0 — the secret shopper: see what a new lead actually gets, and what every workflow step really did
 
 Pick a client, tick the workflows a new lead goes through, and press Start. It makes a clearly marked test lead with your own email address and no first name, adds it to those workflows, and watches until the last planned message is due. Then it reads back every text and email the account actually sent, word for word, and shows what a customer would have seen wrong: a blank name ("Hi , "), template code left showing, another of your accounts' business names, a link that doesn't open, a message sent twice, or a planned message that never came.

package/README.md

@@ -897,7 +897,7 @@ Built by **[Elite DCs, LLC](https://elitedcs.com)** — a digital marketing and
 
 **Tech stack:** TypeScript, Node.js, esbuild, MCP SDK, Zod, GHL API v2, Firebase Auth
 
-**Version:** 3.85.0
+**Version:** 3.87.0
 
 ---