C5.1: Public Surface Governance & Claims Alignment v2.3.7 (#12)

SynapseLayer · web-flow · commit ffdd11b4e8c7 · 2026-05-04T16:39:43.000-03:00
- PUBLIC_CLAIMS_MATRIX.md v1.0 governance foundation
- Removed all FORBIDDEN claims (Zero-Knowledge, Continuous Consciousness, Giving Models a Soul)
- Version sync: server.json + smithery.yaml → 2.3.7
- CHANGELOG entry added
- 20 files updated (copy/metadata only, zero logic changes)
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,14 @@
 # Changelog
 
+## [2.3.7] - 2026-05-04
+
+### Changed
+- Public surface governance: claims matrix v1.0 applied
+- Removed deprecated claims from all public surfaces
+- Aligned descriptions, topics and metadata across all public repos
+- Version synchronized across all distribution channels
+
+
 ## [1.2.0] — 2026-04-21
 
 ### Added
diff --git a/docs/PUBLIC_CLAIMS_MATRIX.md b/docs/PUBLIC_CLAIMS_MATRIX.md
@@ -0,0 +1,68 @@
+# PUBLIC_CLAIMS_MATRIX.md
+**Synapse Layer — Matriz de Claims Públicos v1.0**
+
+Última atualização: 04 de Maio de 2026
+Proprietário: Ismael Marchi
+Status: APPROVED
+
+## Regra de Ouro
+> Público explica o benefício. Privado preserva o moat.
+
+## ALLOWED — Claims Aprovados
+
+### Hero / One-liner
+- "RAG retrieves. Synapse remembers."
+- "Persistent, encrypted memory infrastructure for AI agents."
+
+### Descrição Principal
+- Persistent memory infrastructure for AI agents — encrypted, governed, and cross-agent.
+- Memory that survives across sessions, models, and tools.
+- State Continuity Layer for autonomous AI systems.
+
+### Benefícios (comprováveis)
+- Cross-model handover seguro
+- Privacy-first: PII redaction + AES-256
+- Deterministic recall com Trust Quotient (TQ)
+- Works in 30 seconds
+- Server never sees plaintext
+
+### Categorias Oficiais
+- Memory Infrastructure Layer (principal)
+- State Continuity Layer (secundária)
+- OAuth for AI Memory (terciária)
+
+## HOLD — Só usar com evidência verificável
+- "70% token reduction" → só com dados públicos reais
+- "Zero context loss across 10K+ sessions" → só após validação
+- "Enterprise-ready" → só após casos reais
+- "Self-healing memory" → em desenvolvimento
+
+## FORBIDDEN — Nunca usar publicamente
+- Zero-Knowledge
+- Continuous Consciousness / Consciência Sintética
+- Giving Models a Soul
+- Imortalidade Cognitiva
+- GDPR/LGPD compliant (usar: "designed for LGPD/GDPR alignment")
+- Detalhes de arquitetura interna (TQ weights, scoring, heuristics, pipeline)
+- Métricas sem dados verificáveis
+
+## Copy Aprovada por Canal
+
+| Canal | Copy Aprovada |
+|---|---|
+| GitHub Org Bio | Persistent, encrypted memory infrastructure for AI agents. Server never sees plaintext. MCP-native. Built in Brazil. |
+| synapse-layer description | Persistent memory infrastructure for AI agents — encrypted, governed, and cross-agent. |
+| synapse-sdk-python description | Python SDK for Synapse Layer — persistent, encrypted, cross-agent memory for AI agents. |
+| synapse-layer-skill description | Persistent memory skill for AI agents. Remembers preferences, brand voice and decisions with Trust Quotient scoring. |
+| PyPI / npm | Synapse Layer SDK — persistent, encrypted, cross-agent memory for AI agents. |
+| MCP / Smithery | Persistent memory infrastructure for AI agents. Server never sees plaintext. Works in 30 seconds. |
+
+## Regras de Governança
+1. Toda superfície pública respeita esta matriz.
+2. Novo claim → aprovação antes de publicar.
+3. Se não está em ALLOWED → é HOLD ou FORBIDDEN.
+4. Prioridade: Benefício > Funcionalidade > Como funciona.
+5. Forge herda narrativa do domínio principal. Nunca cria claims paralelas.
+
+Aprovado por: Ismael Marchi
+Próxima revisão: 18 de Maio de 2026
diff --git a/docs/distribution/SMITHERY_SUBMISSION.md b/docs/distribution/SMITHERY_SUBMISSION.md
@@ -10,7 +10,7 @@ Use this content for the Smithery listing page, GitHub PR body, or registry form
 `@synapselayer/synapse-protocol`
 
 ### Display Name
-Synapse Layer — Continuous Consciousness for AI
+Synapse Layer — Persistent Memory for AI Agents
 
 ### One-Liner
 Persistent encrypted memory infrastructure for AI agents. Store, recall, and transfer AES-256-GCM encrypted context across models.
diff --git a/docs/registry/MCP_REGISTRY_SUBMISSION.md b/docs/registry/MCP_REGISTRY_SUBMISSION.md
@@ -5,7 +5,7 @@
 | Field | Value |
 |---|---|
 | **Name** | `io.github.SynapseLayer/synapse-layer` |
-| **Title** | Synapse Layer — Continuous Consciousness Infrastructure |
+| **Title** | Synapse Layer — Persistent Memory Infrastructure |
 | **Version** | `1.1.5` |
 | **License** | Apache-2.0 |
 | **Schema** | `2025-12-11` |
@@ -105,7 +105,7 @@ Persistent encrypted memory for AI agents. AES-256-GCM at rest, PII redaction, H
 {
   "$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",
   "name": "io.github.SynapseLayer/synapse-layer",
-  "title": "Synapse Layer — Continuous Consciousness Infrastructure",
+  "title": "Synapse Layer — Persistent Memory Infrastructure",
   "description": "Persistent encrypted memory for AI agents. AES-256-GCM at rest, PII redaction, HMAC-SHA-256 integrity.",
   "version": "1.1.1",
   "websiteUrl": "https://synapselayer.org",
diff --git a/docs/releases/RELEASE_NOTES_v1.0.6.md b/docs/releases/RELEASE_NOTES_v1.0.6.md
@@ -191,7 +191,7 @@ This release upgrades Synapse Layer from a memory SDK to a **complete Cognitive
 
 *Semantic Privacy Guard™ is now mandatory. Content is AES-256-GCM encrypted at rest.*
 
-**Giving Agents a Past. Giving Models a Soul. ⚗️**
+<!-- Tagline updated: see PUBLIC_CLAIMS_MATRIX.md -->
 ```
 
 ---
diff --git a/examples/crewai_agent.py b/examples/crewai_agent.py
@@ -84,7 +84,7 @@ def main():
     print(f"  Categories: {info.categories}")
     print()
 
-    print("Done. Memory persists across sessions with zero-knowledge security.")
+    print("Done. Memory persists across sessions with encrypted persistent memory.")
 
 
 if __name__ == "__main__":
diff --git a/examples/langchain_agent.py b/examples/langchain_agent.py
@@ -59,7 +59,7 @@ async def main():
     #     config={"configurable": {"session_id": "user-123"}},
     # )
 
-    print("Done. Memory persists across sessions with zero-knowledge security.")
+    print("Done. Memory persists across sessions with encrypted persistent memory.")
 
 
 if __name__ == "__main__":
diff --git a/examples/mcp-secure-memory/README.md b/examples/mcp-secure-memory/README.md
@@ -1,4 +1,4 @@
-# Reference Implementation: Zero-Knowledge Secure Memory for MCP
+# Reference Implementation: Encrypted Persistent Memory for MCP
 
 > **Give any MCP-compatible agent a privacy-first, audit-ready memory layer in under 5 minutes.**
 
@@ -43,7 +43,7 @@ Raw Input ──► PII Sanitization ──► Intent Validation ──► DP No
 | **Intent Validation** | Every memory is classified by intent category with confidence scoring |
 | **Self-Healing** | Conflicting memories with overlapping semantics are automatically reclassified via keyword consensus |
 | **Audit Trail** | Every `store` returns: `memory_id`, `content_hash` (SHA-256), `trust_quotient`, and full sanitization/privacy details |
-| **Zero-Knowledge** | The SDK processes all data locally — no external API calls, no telemetry, no data leaves your environment |
+| **Privacy-First** | The SDK processes all data locally — no external API calls, no telemetry, no data leaves your environment |
 
 ---
 
diff --git a/examples/mcp-secure-memory/server.py b/examples/mcp-secure-memory/server.py
@@ -1,5 +1,5 @@
 """
-Reference Implementation: Zero-Knowledge Secure Memory for MCP (Synapse Layer)
+Reference Implementation: Encrypted Persistent Memory for MCP (Synapse Layer)
 
 A minimal MCP server that exposes two tools — store_memory and recall_memory —
 backed by Synapse Layer's privacy-first memory SDK. Every memory goes through
diff --git a/pyproject.toml b/pyproject.toml
@@ -5,15 +5,15 @@ build-backend = "setuptools.build_meta"
 [project]
 name = "synapse-layer"
 version = "1.2.0"
-description = "Zero-Knowledge Memory Layer for AI Agents — Cognitive Security Protocol with 4 proprietary seals. Giving Agents a Past. Giving Models a Soul."
+description = "Persistent memory infrastructure for AI agents — encrypted, governed, and cross-agent. Server never sees plaintext."
 readme = "README.md"
 license = {text = "Apache 2.0"}
 authors = [{name = "Ismael Marchi", email = "founder.synapselayer@proton.me"}]
 
 keywords = [
   "ai-agents",
   "memory",
-  "zero-knowledge",
+  "persistent-memory",
   "encryption",
   "mcp",
   "agentic-ai",
diff --git a/server.json b/server.json
@@ -2,8 +2,8 @@
   "$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",
   "name": "io.github.SynapseLayer/synapse-layer",
   "title": "Synapse Layer",
-  "description": "Persistent memory infrastructure for AI agents — encrypted, governed, and cross-agent. State Continuity Layer.",
-  "version": "2.3.3",
+  "description": "Persistent memory infrastructure for AI agents \u2014 encrypted, governed, and cross-agent. State Continuity Layer.",
+  "version": "2.3.7",
   "websiteUrl": "https://synapselayer.org",
   "repository": {
     "url": "https://github.com/SynapseLayer/synapse-layer",
@@ -13,7 +13,7 @@
     {
       "registryType": "pypi",
       "identifier": "synapse-layer",
-      "version": "2.3.3",
+      "version": "2.3.7",
       "runtimeHint": "uvx",
       "transport": {
         "type": "stdio"
@@ -28,7 +28,7 @@
         },
         {
           "name": "SYNAPSE_PRIVACY_EPSILON",
-          "description": "Differential privacy epsilon parameter. Lower = stronger privacy. Range: 0.1–2.0.",
+          "description": "Differential privacy epsilon parameter. Lower = stronger privacy. Range: 0.1\u20132.0.",
           "isRequired": false,
           "isSecret": false,
           "default": "0.5"
diff --git a/smithery.yaml b/smithery.yaml
@@ -3,7 +3,7 @@ displayName: "Synapse Layer"
 description: >
   Persistent memory infrastructure for AI agents — encrypted,
   governed, and cross-agent. State Continuity Layer.
-version: "2.3.3"
+version: "2.3.7"
 icon: "🧠"
 sourceUrl: "https://github.com/SynapseLayer/synapse-layer"
 homepage: "https://synapselayer.org"
diff --git a/synapse_layer/__init__.py b/synapse_layer/__init__.py
@@ -1,5 +1,5 @@
 """
-Synapse Layer — Zero-Knowledge Memory Layer for AI Agents
+Synapse Layer — Persistent Memory Infrastructure for AI Agents
 
 This module provides the public API surface matching the package name.
 ``from synapse_layer import ...`` works as expected after
diff --git a/synapse_memory/__init__.py b/synapse_memory/__init__.py
@@ -1,7 +1,7 @@
 """
-Synapse Memory — Zero-Knowledge Memory Layer for AI Agents
+Synapse Memory — Persistent Memory Infrastructure for AI Agents
 
-Giving Agents a Past. Giving Models a Soul.
+Memory that survives across sessions, models, and tools.
 
 Author: Security & Architecture Team @ Synapse Layer
 License: Apache 2.0
diff --git a/synapse_memory/backends/__init__.py b/synapse_memory/backends/__init__.py
@@ -6,14 +6,14 @@
 Available backends:
     - MemoryBackend:  In-memory (default, non-persistent)
     - SqliteBackend:  Zero-config local persistence via sqlite3
-    - ForgeBackend:   Zero-Knowledge cloud via Forge API (AES-256-GCM)
+    - ForgeBackend:   Encrypted cloud via Forge API (AES-256-GCM)
 
 Usage::
 
     from synapse_memory.backends import SqliteBackend
     memory = SynapseMemory(agent_id="my-agent", backend=SqliteBackend())
 
-    # Zero-Knowledge cloud backend
+    # Encrypted cloud backend
     from synapse_memory.backends import ForgeBackend
     backend = ForgeBackend(api_key="sk_connect_...", encryption_key=b"\\x00"*32)
 
diff --git a/synapse_memory/backends/forge_backend.py b/synapse_memory/backends/forge_backend.py
@@ -1,5 +1,5 @@
 """
-Synapse Layer — Forge Backend (True Zero-Knowledge)
+Synapse Layer — Forge Backend (AES-256-GCM Encrypted)
 
 Production-grade backend for Synapse Layer Forge API.
 
@@ -63,7 +63,7 @@
 
 
 class ForgeBackend:
-    """True Zero-Knowledge backend for Synapse Layer Forge API.
+    """AES-256-GCM encrypted backend for Synapse Layer Forge API.
 
     Store encrypts content CLIENT-SIDE (AES-256-GCM) BEFORE the HTTP
     POST — the server NEVER receives or sees plaintext.
@@ -320,7 +320,7 @@ async def store(
         embedding: Optional[List[float]] = None,
         source: str = "sdk",
     ) -> str:
-        """Store a memory via Forge API — True Zero-Knowledge.
+        """Store a memory via Forge API — server never sees plaintext.
 
         Flow:
             1. Encrypt content CLIENT-SIDE (AES-256-GCM)
diff --git a/synapse_memory/client.py b/synapse_memory/client.py
@@ -1,7 +1,7 @@
 """Synapse Layer — Simple Client (sync one-liner API).
 
 Sync HTTP client for Forge API endpoints. Server-side AES-256-GCM
-encryption; the server owns the master key. For client-side zero-knowledge,
+encryption; the server owns the master key. For client-side encrypted storage,
 use :class:`synapse_memory.backends.forge_backend.ForgeBackend` instead.
 
 Usage::
diff --git a/synapse_memory/core.py b/synapse_memory/core.py
@@ -1,5 +1,5 @@
 """
-SynapseMemory — Zero-Knowledge Memory Core with Cognitive Security™
+SynapseMemory — Persistent Memory Core with AES-256-GCM Encryption
 
 Orchestrates the secure memory pipeline with built-in sanitization,
 intent validation, differential privacy, and self-healing recall.
@@ -65,7 +65,7 @@ class RecallResult:
 
 
 class SynapseMemory:
-    """Zero-Knowledge Memory Layer for AI Agents.
+    """Persistent Memory Layer for AI Agents.
 
     Provides persistent, encrypted, cross-model memory with:
     - Mandatory content sanitization (PII removal)
diff --git a/synapse_memory/integrations/langchain_memory.py b/synapse_memory/integrations/langchain_memory.py
@@ -191,7 +191,7 @@ def add_messages(self, messages: Sequence[BaseMessage]) -> None:
     def clear(self) -> None:
         """Clear is a no-op in the OSS adapter.
 
-        Synapse Layer's zero-knowledge architecture does not support
+        Synapse Layer's encrypted architecture does not support
         bulk deletion from the SDK. In production deployments,
         memory lifecycle is managed via retention policies.
         """
diff --git a/tests/test_forge_backend.py b/tests/test_forge_backend.py
@@ -1,7 +1,7 @@
 """
 Synapse Layer — ForgeBackend Unit Tests
 
-Verifies TRUE Zero-Knowledge guarantees:
+Verifies encryption guarantees (server never sees plaintext):
   - TEST-ZK-1: store() encrypts client-side — NO plaintext in HTTP payload
   - TEST-2: recall() decrypts encrypted envelope before returning
   - TEST-3, TEST-4: Auth failures raised correctly
@@ -77,7 +77,7 @@ def _make_encrypted_payload(key: bytes, plaintext: str) -> dict:
 async def test_zk_store_no_plaintext_in_payload(backend: ForgeBackend) -> None:
     """⚡ CRITICAL: Plaintext must NEVER appear in the HTTP payload.
     SDK encrypts client-side BEFORE POST.  Server NEVER sees plaintext.
-    This is THE definitive Zero-Knowledge test.
+    This is THE definitive encryption test — server never sees plaintext.
     """
     secret_content = "My SSN is 123-45-6789 and my password is hunter2"
     captured_body: dict = {}
