@amityco/social-plus-vise 1.4.5 → 1.4.6

package/CHANGELOG.md

@@ -8,6 +8,14 @@ The format is loosely based on [Keep a Changelog](https://keepachangelog.com/en/
 
 No changes yet.
 
+## 1.4.6 — 2026-07-02
+
+### Added
+- **Marketing capability explainer.** Added a shipped `docs/vise-how-it-helps.html` page for marketing and sales enablement that explains Vise's agentic workflow, three validation layers, current real-E2E scorecard methodology, and realistic brownfield-style Social+ integration outputs in a light themed gallery.
+
+### Changed
+- **Preview status is explicit in the public README.** The npm README now marks Vise as a preview release and sets the expected boundary for API stability, benchmark methodology, advisory intelligence, code review, QA, and runtime verification.
+
 ## 1.4.5 — 2026-07-02
 
 ### Changed

package/README.md

@@ -6,6 +6,7 @@
 
 <p align="center">
   <a href="https://www.npmjs.com/package/@amityco/social-plus-vise"><img src="https://img.shields.io/npm/v/@amityco/social-plus-vise.svg" alt="npm version" /></a>
+  <img src="https://img.shields.io/badge/Status-Preview-orange.svg" alt="Preview status" />
   <img src="https://img.shields.io/badge/License-Proprietary-blue.svg" alt="License" />
   <a href="https://learn.social.plus"><img src="https://img.shields.io/badge/Docs-learn.social.plus-informational" alt="Docs" /></a>
 </p>
@@ -33,6 +34,8 @@
 
 ## What is Vise?
 
+> **Preview release.** Vise is usable for guided social.plus SDK integrations, but the CLI surface, evidence format, benchmark methodology, and advisory intelligence may still change as the product hardens through more real projects. Treat it as governed AI assistance with strong validation gates, not a substitute for code review, product QA, or runtime verification.
+
 **Vise wraps your AI coding agent in compliance guardrails while it integrates social.plus SDKs.** The agent still writes the code — Vise grounds that work in real SDK APIs, checks correctness and completeness, helps align generated UI with your design system, and asks you the calls only a human should make. It does not make AI agents perfect; it gives them a governed loop, explicit gates, and evidence when something is green, intentionally scoped out, or still unverified.
 
 While the agent works, Vise:
@@ -153,6 +156,19 @@ Vise is measured against complete brownfield integrations, not isolated snippets
 
 Average with Vise: **91.8**. Average without Vise: **48.0**. Average improvement: **+43.8 points**.
 
+**How the score is calculated:** each run is scored from 0-100 across five dimensions, then weighted into one overall score: capability activation (**25%**), product outcome (**25%**), design integration (**20%**), correctness (**20%**), and efficiency (**10%**). In shorthand:
+
+```text
+overall =
+  capability activation * 0.25 +
+  product outcome       * 0.25 +
+  design integration    * 0.20 +
+  correctness           * 0.20 +
+  efficiency            * 0.10
+```
+
+Capability activation measures whether the Vise arm actually used the governed workflow (inspection, Engagement Intelligence selection, blueprint/sign-off artifacts, docs lookup, checks, sensors, and local `sp-vise/` evidence); no-Vise controls intentionally score 0 on that dimension because they must not use Vise. Product outcome and design integration are judged from runtime-visible social surfaces and screenshots. Correctness covers build/runtime hygiene, SDK-backed source, no leaked secrets, no hardcoded social object IDs, and completed handoff. Efficiency rewards validated, SDK-backed outcome delivered within the time/command budget. Caps lower the score when a run is quota-limited, explicitly unfinished, missing a required multi-surface workplan, or proves too little runtime surface coverage. Platform rows enter the head-to-head average only when the Vise and no-Vise arms target the same platform/scenario and are scored with the same launch-verification standard.
+
 What changed with Vise in the loop:
 
 - The agent starts from a grounded plan instead of guessing the integration shape.