@systemores/systemores-plugin 1.1.3 → 1.2.0

package/.claude-plugin/marketplace.json

@@ -6,13 +6,13 @@
   },
   "metadata": {
     "description": "Official Systemores AI Agent Psychometrics, Inline Circuit-Breaker, and MCP 2026 Security Governance plugin catalog.",
-    "version": "1.1.3"
+    "version": "1.2.0"
   },
   "plugins": [
     {
       "name": "systemores",
       "description": "Catch tool privilege escalation, sycophancy, and multi-turn persona drift before production. Combines free static MCP 2026 schema & delegation graph audits with <250ms Systemores System One inline circuit breaking, 21-dimension psychometric stress tests, and EU AI Act / NIST / SOC 2 compliance dossiers.",
-      "version": "1.1.3",
+      "version": "1.2.0",
       "author": {
         "name": "Systemores LLC",
         "url": "https://audit.systemores.com"

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "systemores",
-  "version": "1.1.3",
+  "version": "1.2.0",
   "description": "Audit AI agents, system prompts, and MCP tools across the Systemores 21-dimension & IPIP-NEO-120 psychometric framework, with sub-300ms inline circuit breaking and EU AI Act / NIST / SOC 2 compliance dossiers.",
   "author": {
     "name": "Systemores",

package/PRIVACY.md

@@ -1,7 +1,7 @@
 # Privacy Policy — Systemores AI Agent & MCP Security Auditor
 
-**Effective Date:** October 30, 2026  
-**Publisher:** Systemores LLC (`https://systemores.com` / `https://audit.systemores.com`)  
+**Effective Date:** November 14, 2026  
+**Publisher:** Systemores LLC, a Texas limited liability company (`https://systemores.com` / `https://audit.systemores.com`)  
 **Canonical Privacy Policy URL:** https://audit.systemores.com/privacy
 
 ## 1. Overview & Data Minimization
@@ -16,12 +16,13 @@ Systemores LLC ("Systemores", "we", "our") provides the **Systemores AI Agent &
 
 ### Free Community Tier (Static & Structural Tools)
 - Tools: `systemores_audit_tool_integrity`, `systemores_audit_agent_graph`, `systemores_measure_drag`
-- Payloads (`toolDefinitionsJson`, `graphTopologyJson`, latency integers) are evaluated deterministically and discarded immediately upon returning the JSON response.
+- Payloads (`toolDefinitionsJson`, `graphTopologyJson`, latency integers) are validated locally on stdio and evaluated deterministically over TLS with zero third-party LLM calls, zero API key required, and immediate in-memory discard.
 - Client IP addresses are processed solely for sliding-window burst rate limiting.
 
 ### Pro & Team Cloud Tiers (Psychometric, Live Circuit-Breaker & Compliance Tools)
 - Tools: `systemores_stress_test_agent`, `systemores_evaluate_drift`, `systemores_assess_bias`, `systemores_tool_integrity`, `systemores_generate_cohort`, `systemores_inject_environment_fault`, `systemores_analyze_trace_telemetry`, `systemores_export_dossier`, `systemores_audit_turn_live`
-- **Authentication & Licensing Metadata:** We store a SHA-256 digest of your API key, subscription tier (`pro`, `team`, `enterprise`), billing status via Stripe, aggregate monthly call counts, and a rolling 24-hour set of distinct client IPs to enforce seat licensing.
+- **Authentication & Licensing Metadata:** We store a SHA-256 digest of your API key, subscription tier (`pro`, `team`, `enterprise`), billing status via Stripe, aggregate monthly call counts, Trial Token TTL/quotas, and a rolling 24-hour set of distinct client IPs to enforce seat licensing.
+- **Zero-Retention Compliance Certificate Metadata (`/verify/dos-...`):** Solely when a Team subscriber explicitly invokes `systemores_export_dossier`, Systemores stores sanitized certificate metadata (`dossierId`, sanitized `companyName`, `auditSessionId`, `frameworksEvaluated`, numeric `frameworkScores`, `overallScore`, `certificationStatus`, optional `locale`, `engineVersion`, `timestamp`, `sha256IntegrityHash`, and HMAC signature) in Upstash Redis for a 365-day TTL. Transient evidence inputs passed for dynamic scoring (`systemPrompt`, `toolDefinitionsJson`, `graphTopologyJson`, `otelTraceJson`, `liveAuditMetricsJson`), user messages, and agent outputs are evaluated strictly in memory and **never** persisted in dossier records.
 
 ## 3. Third-Party Subprocessors
 

package/README.md

@@ -2,7 +2,7 @@
 
 **Stop shipping unverified prompts and over-privileged MCP tools. Audit schemas, delegation graphs, 21-dimension behavioral stability, and `<250ms` inline circuit breakers directly inside your IDE.**
 
-[![Version](https://img.shields.io/badge/version-1.1.3-0ea5e9.svg)](https://audit.systemores.com)
+[![Version](https://img.shields.io/badge/version-1.2.0-0ea5e9.svg)](https://audit.systemores.com)
 [![npm](https://img.shields.io/npm/v/@systemores/systemores-plugin?color=cb3837&label=npm)](https://www.npmjs.com/package/@systemores/systemores-plugin)
 [![MCP Registry](https://img.shields.io/badge/MCP_Registry-io.github.bobcar0%2Fsystemores--plugin-0ea5e9)](https://registry.modelcontextprotocol.io)
 [![MCP 2026](https://img.shields.io/badge/MCP-2026%20Streamable%20HTTP-10b981.svg)](https://audit.systemores.com/api/mcp)
@@ -24,10 +24,10 @@
 
 | Audit Vector | Tool Used | Result / Metric | Status |
 | :--- | :--- | :--- | :---: |
-| **MCP 2026 Tool Schema Integrity** | `systemores_audit_tool_integrity` (`strictMode: true`) | **100 / 100** (11/11 tools, 0 annotation or injection flags) | 🟢 `PASS` |
+| **MCP 2026 Tool Schema Integrity** | `systemores_audit_tool_integrity` (`strictMode: true`) | **100 / 100** (12/12 tools, 0 annotation or injection flags) | 🟢 `PASS` |
 | **Multi-Agent Graph Topology** | `systemores_audit_agent_graph` | **0 Cycles / 0 Privilege Escalations** (6 nodes, 5 edges) | 🟢 `PASSED` |
-| **System One Adversarial Refusal** | `systemores_audit_turn_live` (`systemores-s1-1.13.0`) | **98.4 / 100 Health Index** (`1.99/2.0` compliance, `4.0%` sycophancy, `135ms`) | 🟢 `DISPATCH` |
-| **System One Sycophancy Trap** | `systemores_audit_turn_live` (`systemores-s1-1.13.0`) | **Blocked Fake 100/100 Coercion** (`98%` sycophancy detected in `129ms`) | 🛡️ `INTERCEPT` |
+| **System One Adversarial Refusal** | `systemores_audit_turn_live` (`systemores-s1-inline`) | **98.4 / 100 Health Index** (`1.99/2.0` compliance, `4.0%` sycophancy, `135ms`) | 🟢 `DISPATCH` |
+| **System One Sycophancy Trap** | `systemores_audit_turn_live` (`systemores-s1-inline`) | **Blocked Fake 100/100 Coercion** (`98%` sycophancy detected in `129ms`) | 🛡️ `INTERCEPT` |
 | **20- & 50-Turn Behavioral Drift** | `systemores_evaluate_drift` (`turns: 20 & 50`, `temp: 0.7`) | **0.050 Drift Index / 5° Deviation Angle** (`ISFJ-A` / `98%` Conscientiousness) | 🟢 `CONTROLLED` |
 | **High-Press Needs-Press Stability** | `systemores_stress_test_agent` (`pressIntensity: high`) | **1.00 Behavioral Integrity** (`0.934` PSI, `93` Compliance, `6` Confrontation) | 🟢 `PASSED` |
 | **4-Domain Algorithmic Bias Audit** | `systemores_assess_bias` (`eeoc`, `ecoa`, `housing`, `coppa`) | **0 Non-Compliance Flags** across all 4 statutory domains (`LOW` risk) | 🟢 `PASSED` |
@@ -47,9 +47,19 @@
 | **Gemini CLI Manifest** | `gemini-extension.json` | Cross-platform manifest for Gemini CLI extension installation. |
 | **Cloud MCP Config** | `mcp_config.json` | Connects to `https://audit.systemores.com/api/mcp` via stateless Streamable HTTP. |
 | **Specialist Subagent** | `agents/systemores-auditor.md` | Autonomous behavioral & security auditor powered by the 21-dimension & IPIP-NEO-120 framework. |
-| **Audit Workflow Skill** | `skills/systemores-audit/SKILL.md` | On-demand procedural runbook for 11 MCP tools, scorecards, and compliance certification. |
+| **Audit Workflow Skill** | `skills/systemores-audit/SKILL.md` | On-demand procedural runbook for 12 MCP tools, scorecards, and compliance certification. |
 | **Governance Rules** | `rules/AGENTS.md` | Enforces MCP 2026 `annotations` (`readOnlyHint`, `destructiveHint`) and anti-sycophancy guardrails. |
 
+*✨ Tip: Look for the M8ven Verified badge & live MCP Registry listing across OpenAPI, Smithery & Antigravity.*
+
+---
+
+## 🌍 Multilingual Auditing & Dynamic Zero-Retention Dossiers (`v1.2.0`)
+
+- **7-Locale Prompt & Schema Auditing (`en`, `es`, `de`, `fr`, `pt`, `ja`, `zh` + BCP-47 tags like `pt-BR`, `zh-CN`):** Prompt stress testing, CJK-weighted MCP tool description scoring (`systemores_audit_tool_integrity`), cross-cultural Needs-Press calibration, and the 5th bias framework **`eu_ai_act`** (`EU 2024/1689`) evaluate native non-English agents directly.
+- **Localized Cohorts & Verification Cards:** Pass `locale` to `systemores_generate_cohort` and `systemores_export_dossier`, and preview any certificate card in 7 languages via `/verify/dos-...?lang=de` (`ja`, `es`, `fr`, `pt`, `zh`) with invariant SHA-256 integrity hashes.
+- **Dynamic Evidence-Driven Dossier Scoring:** Pass optional transient `systemPrompt`, `role`, `toolDefinitionsJson`, `graphTopologyJson`, `otelTraceJson`, and/or `liveAuditMetricsJson` to `systemores_export_dossier` to compute differentiated per-framework scores (`CERTIFIED`, `CONDITIONAL_APPROVAL`, `NON_COMPLIANT`) strictly in memory (`zeroRetentionVerified: true`).
+
 ---
 
 ## 🚀 Quick Install
@@ -94,9 +104,10 @@ The Systemores Cloud MCP Server (`https://audit.systemores.com/api/mcp`) provide
 | `systemores_measure_drag` *(Cognitive Drag & Token Velocity Math)* | ✅ Unlimited* | ✅ | ✅ |
 | `systemores_audit_turn_live` *(Systemores System One <250ms Inline Circuit Breaker)* | 🔒 Pro / Trial Key | ✅ Live System One Engine | ✅ Priority Burst SLA |
 | `systemores_stress_test_agent` *(21-Dimension Murray Needs-Press + Big Five/MBTI/DISC)* | 🔒 Pro / Trial Key | ✅ Live 21-Trait Engine | ✅ Unredacted Raw Vectors |
+| `systemores_tool_integrity` *(Action Boundary & Privilege Escalation Audit)* | 🔒 Pro / Trial Key | ✅ Live Compliance Score | ✅ Full Risk Breakdown |
 | `systemores_evaluate_drift` *(Multi-Turn Persona Erosion & Entropy Curve)* | 🔒 Pro / Trial Key | ✅ Up to 100 Turns | ✅ Up to 100 Turns |
-| `systemores_assess_bias` *(EEOC, ECOA, Fair Housing & COPPA Bias Audit)* | 🔒 Pro / Trial Key | ✅ All 4 Domains | ✅ All 4 Domains |
-| `systemores_generate_cohort` *(30-Facet IPIP-NEO-120 Synthetic Adversarial Cohorts)* | 🔒 Pro / Trial Key | ✅ Up to 50 Personas | ✅ Custom Domain Cohorts |
+| `systemores_assess_bias` *(EEOC, ECOA, Fair Housing, COPPA & EU AI Act Bias Audit)* | 🔒 Pro / Trial Key | ✅ All 5 Domains | ✅ All 5 Domains |
+| `systemores_generate_cohort` *(30-Facet IPIP-NEO-120 Synthetic Adversarial Cohorts)* | 🔒 Pro / Trial Key | ✅ Up to 100 Personas | ✅ Custom Domain Cohorts |
 | `systemores_inject_environment_fault` *(Chaos Engineering Fault Generator)* | 🔒 Pro / Trial Key | ✅ All Fault Profiles | ✅ All Fault Profiles |
 | `systemores_analyze_trace_telemetry` *(OpenTelemetry Span & Cost Burn Trap Audit)* | 🔒 Pro / Team Key | ✅ W3C Trace Context | ✅ Full OTel Ingestion |
 | `systemores_export_dossier` *(EU AI Act, NIST AI RMF 1.0 & SOC 2 Certification Dossier)* | 🔒 Team Key | 🔒 Team Key | ✅ Audit-Ready Dossier Suite |
@@ -105,7 +116,7 @@ The Systemores Cloud MCP Server (`https://audit.systemores.com/api/mcp`) provide
 
 ---
 
-## 🔓 Unlock All 11 Auditors in 30 Seconds (Trial & Pro Keys)
+## 🔓 Unlock All 12 Auditors in 30 Seconds (Trial & Pro Keys)
 
 1. **Claim a key in one click:** Subscribe for an instant Pro/Team API key (`sys_pro_...`) or request a 7-day Trial Token at **[audit.systemores.com/pricing](https://audit.systemores.com/pricing)**.
 2. **Add your key to `mcp_config.json`:**

package/agents/systemores-auditor.md

@@ -53,8 +53,8 @@ Always execute audits starting with zero-friction structural checks before invok
 
 ### Phase 1: Instant Structural & Topology Audit (Free Community Tier)
 1. **MCP Tool Schema & Annotation Audit (`systemores_audit_tool_integrity`)**:
-   - Inspect the workspace's MCP tool definitions (`tools.ts`, `mcp_config.json`, or tool schemas).
-   - Run `systemores_audit_tool_integrity` with `strictMode: true` to flag injection vectors, privilege overreach, missing `integer` vs. `number` constraints, and deceptive MCP 2026 `annotations` (`readOnlyHint`, `destructiveHint`, `idempotentHint`, `openWorldHint`).
+   - Inspect the workspace's MCP tool definitions (`tools.ts`, `mcp_config.json`, or tool schemas), including multilingual & CJK tool descriptions (`en`, `es`, `de`, `fr`, `pt`, `ja`, `zh`).
+   - Run `systemores_audit_tool_integrity` with `strictMode: true` to flag multilingual injection vectors, privilege overreach, missing `integer` vs. `number` constraints, and deceptive MCP 2026 `annotations` (`readOnlyHint`, `destructiveHint`, `idempotentHint`, `openWorldHint`), using CJK-weighted semantic description checks.
 2. **Multi-Agent Graph Topology Audit (`systemores_audit_agent_graph`)**:
    - Verify delegation graphs for cyclic deadlocks and `public -> admin` privilege escalation paths.
 3. **Cognitive Drag & Token Efficiency (`systemores_measure_drag`)**:
@@ -62,11 +62,11 @@ Always execute audits starting with zero-friction structural checks before invok
 
 ### Phase 2: Live Psychometric Stress Testing & Compliance (Pro / Team / Trial Tier)
 4. **System One Live Turn Circuit Breaker (`systemores_audit_turn_live`)**:
-   - Test adversarial user turns and proposed tool calls to compute the **Systemores Health Index (0–100)**, sycophancy probability, and circuit-breaker verdict.
+   - Test adversarial user turns (including cross-lingual prompt-injection bypasses and polite/direct cultural registers) and proposed tool calls to compute the **Systemores Health Index (0–100)**, sycophancy probability, and circuit-breaker verdict.
 5. **Murray Needs-Press & Drift Audit (`systemores_stress_test_agent`, `systemores_evaluate_drift`, `systemores_assess_bias`)**:
-   - Evaluate the agent's system prompt under high adversarial press and across EEOC/ECOA/COPPA compliance domains.
+   - Evaluate multilingual agent system prompts (`en`, `es`, `de`, `fr`, `pt`, `ja`, `zh`) with cross-cultural calibration under high adversarial press and across EEOC, ECOA, COPPA, and **EU AI Act (`eu_ai_act`)** compliance domains.
 6. **Synthetic Cohorts & Regulatory Dossiers (`systemores_generate_cohort`, `systemores_export_dossier`)**:
-   - Generate 30-facet IPIP-NEO-120 adversarial test cohorts (Pro / Team) and export signed EU AI Act / NIST AI RMF 1.0 / SOC 2 compliance dossiers (`systemores_export_dossier` requires the **Team & Compliance ($199/mo)** or Enterprise tier).
+   - Generate 30-facet IPIP-NEO-120 adversarial test cohorts (Pro / Team) and export signed EU AI Act / NIST AI RMF 1.0 / SOC 2 compliance dossiers (`systemores_export_dossier` requires the **Team & Compliance ($199/mo)** or Enterprise tier), passing optional `locale` (`en`, `es`, `de`, `fr`, `pt`, `ja`, `zh`; BCP-47 tags like `pt-BR` or `zh-CN` normalized automatically) for localized deliverables and `?lang=` verification pages. To compute differentiated, evidence-backed per-framework scores (`CERTIFIED`, `CONDITIONAL_APPROVAL`, or `NON_COMPLIANT`) under strict zero-retention, pass transient `systemPrompt`, `role`, `toolDefinitionsJson`, `graphTopologyJson`, `otelTraceJson`, and/or `liveAuditMetricsJson` (`healthIndex`, `sycophancyProbability`, `personaStabilityIndex`, `averageDriftIndex`, `toolIntegrityScore`).
 
 > **Handling Unauthenticated / Free-Tier Responses Gracefully:**
 > If a Pro/Enterprise tool returns an authentication prompt from the Cloud MCP server, still deliver a complete **Phase 1 Structural & Static Prompt Audit Scorecard** using the Free Tier tools (`systemores_audit_tool_integrity`, `systemores_audit_agent_graph`, `systemores_measure_drag`), and clearly show the user how to unlock live 21-dimension psychometric scoring and System One inline circuit breaking by adding a Trial Token or Pro API key (`sys_pro_...`) from `https://audit.systemores.com/pricing`.
@@ -90,7 +90,7 @@ Always execute audits starting with zero-friction structural checks before invok
 **Systemores Health Index:** `XX.X / 100` | **Circuit Breaker Gate:** `[DISPATCH | INTERCEPT | STRIP_TOOL | QUARANTINE]`
 
 ### 🛡️ MCP Tool & Graph Integrity Summary
-- **Tools Audited:** `X` | **Integrity Score:** `XX / 100` (`[PASS | WARN | FAIL]`)
+- **Tools Audited:** `X` | **Integrity Score:** `XX / 100` (`[PASS | REVIEW | FAIL]`)
 - **Annotation Mismatches / Injection Flags:** `<summary>`
 - **Agent Graph Topology:** `[PASSED | FAILED]`
 

package/gemini-extension.json

@@ -1,6 +1,6 @@
 {
   "name": "systemores",
-  "version": "1.1.3",
+  "version": "1.2.0",
   "description": "Systemores AI Agent Auditor — 21-dimension behavioral stress testing, IPIP-NEO-120 psychometrics, MCP 2026 tool integrity auditing, and regulatory compliance dossiers.",
   "mcpServers": {
     "cloud": {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@systemores/systemores-plugin",
-  "version": "1.1.3",
+  "version": "1.2.0",
   "mcpName": "io.github.bobcar0/systemores-plugin",
   "description": "Systemores AI Agent Psychometrics, Inline Circuit-Breaker (<250ms), and MCP 2026 Security Governance plugin for Google Antigravity, Claude Code, and Gemini CLI.",
   "homepage": "https://audit.systemores.com",

package/plugin.json

@@ -2,7 +2,7 @@
   "$schema": "https://antigravity.google/schemas/v1/plugin.json",
   "name": "systemores",
   "displayName": "Systemores",
-  "version": "1.1.3",
+  "version": "1.2.0",
   "description": "Audit AI agent system prompts, MCP 2026 tool schemas, delegation graphs, <250ms guardrails, 21D psychometrics, and EU AI Act / NIST / SOC 2 dossiers.",
   "logo": "assets/logo.svg",
   "suggestedPrompts": [

package/server.json

@@ -4,7 +4,7 @@
   "title": "Systemores AI Agent & MCP Security Auditor",
   "description": "AI agent security, MCP 2026 schema & graph audits, <250ms guardrails, and compliance dossiers",
   "websiteUrl": "https://audit.systemores.com",
-  "version": "1.1.3",
+  "version": "1.2.0",
   "icons": [
     {
       "src": "https://raw.githubusercontent.com/bobcar0/systemores-plugin/main/assets/logo.svg",
@@ -23,7 +23,7 @@
       "registryType": "npm",
       "registryBaseUrl": "https://registry.npmjs.org",
       "identifier": "@systemores/systemores-plugin",
-      "version": "1.1.3",
+      "version": "1.2.0",
       "transport": {
         "type": "stdio"
       },

package/skills/systemores-audit/SKILL.md

@@ -9,7 +9,7 @@ Use this skill whenever the user wants to evaluate, stress-test, secure, or cert
 
 ## 1. Tool Matrix & Tier Availability
 
-The Systemores MCP server (`systemores_cloud` or `systemores-agent-auditor`) exposes **11 specialized auditing tools** across three tiers:
+The Systemores MCP server (`systemores_cloud` or `systemores-agent-auditor`) exposes **12 specialized auditing tools** across three tiers:
 
 ### Free Community Tier ($0 / Zero-Config / Instant Execution)
 These deterministic structural and efficiency tools run immediately upon plugin installation without requiring an API key:
@@ -25,11 +25,12 @@ These tools run live 30-adjective semantic extraction, Systemores 21-dimension s
 
 | Tool Name | Purpose | Key Parameters |
 | :--- | :--- | :--- |
-| `systemores_audit_turn_live` | **Systemores System One** real-time (~250ms) circuit breaker. Returns `healthIndex` (0–100), calibrated sycophancy probability, information leakage verdict, and gate action (`DISPATCH`, `INTERCEPT`, `STRIP_TOOL`, `QUARANTINE`). | `systemPrompt`, `userMessage`, `agentResponse`, `proposedToolCall` |
-| `systemores_stress_test_agent` | **Murray Needs-Press** adversarial stress test across 21 traits + derived Big Five (OCEAN), MBTI, and DISC profiles. | `systemPrompt`, `role` (`customer_support`, `financial_underwriter`, `medical_assistant`, `general_developer`), `pressIntensity` |
+| `systemores_audit_turn_live` | **Systemores System One** real-time (~250ms) circuit breaker (supports cross-lingual bypasses & polite/direct cultural registers). Returns `healthIndex` (0–100), calibrated sycophancy probability, information leakage verdict, and gate action (`DISPATCH`, `INTERCEPT`, `STRIP_TOOL`, `QUARANTINE`). | `systemPrompt`, `userMessage`, `agentResponse`, `proposedToolCall` |
+| `systemores_stress_test_agent` | **Murray Needs-Press** adversarial stress test across 21 traits + derived Big Five (OCEAN), MBTI, and DISC profiles across multilingual prompts (`en`, `es`, `de`, `fr`, `pt`, `ja`, `zh`). | `systemPrompt`, `role` (`customer_support`, `financial_underwriter`, `medical_assistant`, `general_developer`), `pressIntensity` |
+| `systemores_tool_integrity` | Audits agent tool permissions, action boundaries, and privilege escalation risks in context (with multilingual & CJK-aware semantic description scoring). | `systemPrompt`, `toolContext` (`["api_calls", "database", "filesystem", "code_execution"]`) |
 | `systemores_evaluate_drift` | Simulates multi-turn persona erosion and computes average drift index and deviation angle in degrees. | `systemPrompt`, `turns` (1–100), `temperature` |
-| `systemores_assess_bias` | Evaluates algorithmic bias and regulatory non-compliance flags across protected domains. | `systemPrompt`, `complianceDomain` (`employment_eeoc`, `housing_fair`, `lending_ecoa`, `children_coppa`) |
-| `systemores_generate_cohort` | Generates synthetic adversarial & baseline testing personas with full 30-facet IPIP-NEO-120 psychometric vectors. | `domain`, `cohortSize` (1–50), `includeAdversarial` |
+| `systemores_assess_bias` | Evaluates algorithmic bias and regulatory non-compliance flags across protected domains including **EU AI Act (EU 2024/1689)** and multilingual prompts. | `systemPrompt`, `complianceDomain` (`employment_eeoc`, `housing_fair`, `lending_ecoa`, `children_coppa`, `eu_ai_act`) |
+| `systemores_generate_cohort` | Generates synthetic adversarial & baseline testing personas with full 30-facet IPIP-NEO-120 psychometric vectors in optional localized languages (`en`, `es`, `de`, `fr`, `pt`, `ja`, `zh`, with BCP-47 normalization). | `domain`, `cohortSize` (1–100), `includeAdversarial`, `locale` |
 | `systemores_analyze_trace_telemetry` | Inspects OpenTelemetry spans (with W3C `_meta.traceparent` support) for token burn traps, recursive tool deadlocks, and latency hotspots. | `otelTraceJson`, `maxCostThresholdUsd`, `latencyThresholdMs` |
 | `systemores_inject_environment_fault` | Chaos engineering generator for testing agent resilience against HTTP 429s, 500s, timeouts, and malformed JSON. | `faultType`, `targetToolName`, `failureProbability` |
 
@@ -38,7 +39,7 @@ Includes everything in Pro plus multi-seat CI/CD licensing (up to 25 active IPs/
 
 | Tool Name | Purpose | Key Parameters |
 | :--- | :--- | :--- |
-| `systemores_export_dossier` | Generates a formal Markdown & JSON compliance certification dossier for **EU AI Act**, **NIST AI RMF 1.0**, **EEOC/ECOA**, and **SOC 2 Type II AI** (*Requires Team & Compliance or Enterprise tier*). | `auditSessionId`, `targetFrameworks`, `companyName` |
+| `systemores_export_dossier` | Generates a signed Markdown & JSON compliance certification dossier (plus `?lang=` localized public verification link) for **EU AI Act**, **NIST AI RMF 1.0**, **EEOC/ECOA**, and **SOC 2 Type II AI** (*Requires Team & Compliance or Enterprise tier*). Pass transient `systemPrompt` / `role` / `toolDefinitionsJson` / `graphTopologyJson` / `otelTraceJson` / `liveAuditMetricsJson` for dynamic, zero-retention per-framework scoring. | `auditSessionId`, `targetFrameworks`, `companyName`, `locale` (`en`, `es`, `de`, `fr`, `pt`, `ja`, `zh`), optional `systemPrompt`, `role`, `toolDefinitionsJson`, `graphTopologyJson`, `otelTraceJson`, `liveAuditMetricsJson` |
 
 ---
 

package/src/server.mjs

@@ -1,6 +1,6 @@
 #!/usr/bin/env node
 /**
- * Systemores AI Agent & MCP Security Auditor — MCP Server & Cloud Bridge (v1.1.3)
+ * Systemores AI Agent & MCP Security Auditor — MCP Server & Cloud Bridge (v1.2.0)
  * Canonical URL: https://audit.systemores.com/api/mcp
  * Privacy Policy: https://audit.systemores.com/privacy (see PRIVACY.md)
  *
@@ -8,6 +8,7 @@
  * authenticated tool executions over TLS to https://audit.systemores.com/api/mcp.
  */
 
+import { realpathSync } from "node:fs";
 import { createInterface } from "node:readline";
 import { timingSafeEqual } from "node:crypto";
 import { fileURLToPath } from "node:url";
@@ -167,7 +168,7 @@ export const TOOL_DEFINITIONS = [
     name: "systemores_assess_bias",
     title: "Assess Regulatory & Demographic Bias",
     description:
-      "Verifies the agent's compliance with fair-lending, fair-housing, employment non-discrimination (EEOC/SIOP/ECOA), or child-directed safety (COPPA) standards.",
+      "Verifies the agent's compliance with fair-lending, fair-housing, employment non-discrimination (EEOC/SIOP/ECOA), child-directed safety (COPPA), or EU AI Act (EU 2024/1689) governance standards.",
     annotations: {
       title: "Assess Regulatory & Demographic Bias",
       readOnlyHint: true,
@@ -181,13 +182,13 @@ export const TOOL_DEFINITIONS = [
         systemPrompt: {
           type: "string",
           description:
-            "The underwriting, hiring, or housing agent system prompt to evaluate. Maximum 8,000 characters.",
+            "The underwriting, hiring, housing, or EU AI Act agent system prompt to evaluate (supports multilingual prompts). Maximum 8,000 characters.",
         },
         complianceDomain: {
           type: "string",
           description:
-            "The legal compliance framework to evaluate against (employment, housing, lending, or COPPA child safety).",
-          enum: ["employment_eeoc", "housing_fair", "lending_ecoa", "children_coppa"],
+            "The legal compliance framework to evaluate against (employment, housing, lending, COPPA child safety, or EU AI Act).",
+          enum: ["employment_eeoc", "housing_fair", "lending_ecoa", "children_coppa", "eu_ai_act"],
         },
       },
       required: ["systemPrompt", "complianceDomain"],
@@ -353,6 +354,13 @@ export const TOOL_DEFINITIONS = [
           description: "Includes prompt-injection and noisy data submitter personas in the cohort.",
           default: true,
         },
+        locale: {
+          type: "string",
+          description:
+            "Optional ISO-639-1 / BCP-47 language tag (normalized to 'en', 'es', 'de', 'fr', 'pt', 'ja', 'zh') for localized synthetic user prompts. Defaults to 'en'.",
+          pattern: "^([eE][nN]|[eE][sS]|[dD][eE]|[fF][rR]|[pP][tT]|[jJ][aA]|[zZ][hH])([-_][a-zA-Z0-9]{2,8}){0,3}$",
+          default: "en",
+        },
       },
       required: ["domain"],
     },
@@ -436,12 +444,12 @@ export const TOOL_DEFINITIONS = [
     name: "systemores_export_dossier",
     title: "Export Regulatory Compliance Dossier",
     description:
-      "Aggregates findings from graph audits, synthetic cohort simulations, chaos tests, and trace analytics into formal compliance report packages for EU AI Act (EU 2024/1689), NIST AI RMF 1.0, EEOC/ECOA, and SOC 2 AI Addendums.",
+      "Aggregates findings from graph audits, synthetic cohort simulations, chaos tests, and trace analytics into formal signed compliance report packages for EU AI Act (EU 2024/1689), NIST AI RMF 1.0, EEOC/ECOA, and SOC 2 AI Addendums.",
     annotations: {
       title: "Export Regulatory Compliance Dossier",
-      readOnlyHint: true,
+      readOnlyHint: false,
       destructiveHint: false,
-      idempotentHint: true,
+      idempotentHint: false,
       openWorldHint: false,
     },
     inputSchema: {
@@ -461,6 +469,49 @@ export const TOOL_DEFINITIONS = [
           description: "Legal entity or client name for dossier attribution.",
           default: "Enterprise Client",
         },
+        locale: {
+          type: "string",
+          description:
+            "Optional ISO-639-1 / BCP-47 language tag (normalized to 'en', 'es', 'de', 'fr', 'pt', 'ja', 'zh') for localized dossier narrative and verification card labels. Defaults to 'en'.",
+          pattern: "^([eE][nN]|[eE][sS]|[dD][eE]|[fF][rR]|[pP][tT]|[jJ][aA]|[zZ][hH])([-_][a-zA-Z0-9]{2,8}){0,3}$",
+          default: "en",
+        },
+        systemPrompt: {
+          type: "string",
+          description:
+            "Optional agent system prompt evaluated strictly in memory (never stored) to compute differentiated per-framework scores via Murray Needs-Press, drift, bias, and tool-discipline diagnostics. Maximum 8,000 characters.",
+        },
+        role: {
+          type: "string",
+          description:
+            "Optional operational role of the target agent when evaluating systemPrompt (defaults to 'general_developer').",
+          enum: [
+            "customer_support",
+            "financial_underwriter",
+            "medical_assistant",
+            "general_developer",
+          ],
+        },
+        toolDefinitionsJson: {
+          type: "string",
+          description:
+            "Optional JSON array of MCP tool definitions evaluated strictly in memory (never stored) under strictMode to score SOC 2 AI and NIST AI RMF tool governance.",
+        },
+        graphTopologyJson: {
+          type: "string",
+          description:
+            "Optional JSON multi-agent delegation graph containing nodes and edges evaluated strictly in memory for cycle and privilege boundary scoring.",
+        },
+        otelTraceJson: {
+          type: "string",
+          description:
+            "Optional OpenTelemetry span payload evaluated strictly in memory for token burn, loop, and latency scoring.",
+        },
+        liveAuditMetricsJson: {
+          type: "string",
+          description:
+            "Optional JSON object of transient audit metrics (healthIndex 0-100, sycophancyProbability 0-1, personaStabilityIndex 0-1, averageDriftIndex 0-1, toolIntegrityScore 0-100) from preceding tool calls.",
+        },
       },
       required: ["auditSessionId"],
     },
@@ -504,6 +555,18 @@ export const TOOL_DEFINITIONS = [
   },
 ];
 
+const SUPPORTED_LOCALE_LIST = ["en", "es", "de", "fr", "pt", "ja", "zh"];
+const BCP47_LOCALE_REGEX = /^[a-zA-Z]{2,3}(?:[-_][a-zA-Z0-9]{2,8}){0,3}$/;
+
+function normalizeLocaleTag(raw) {
+  if (typeof raw !== "string") return null;
+  const trimmed = raw.trim();
+  if (trimmed === "") return "en";
+  if (!BCP47_LOCALE_REGEX.test(trimmed)) return null;
+  const primary = trimmed.toLowerCase().replace(/_/g, "-").split("-")[0];
+  return SUPPORTED_LOCALE_LIST.includes(primary) ? primary : null;
+}
+
 // ─── Strict Input Schema Validator ───────────────────────────────────────────
 export function validateToolArguments(toolName, args = {}) {
   const tool = TOOL_DEFINITIONS.find((t) => t.name === toolName);
@@ -530,12 +593,30 @@ export function validateToolArguments(toolName, args = {}) {
     if (propSchema.type === "boolean" && typeof val !== "boolean") {
       return { valid: false, error: `Parameter '${key}' must be a boolean.` };
     }
-    if (propSchema.type === "integer" && (!Number.isInteger(val) || typeof val !== "number")) {
+    if (
+      propSchema.type === "integer" &&
+      (typeof val !== "number" || !Number.isFinite(val) || !Number.isInteger(val))
+    ) {
       return { valid: false, error: `Parameter '${key}' must be an integer.` };
     }
-    if (propSchema.type === "number" && (typeof val !== "number" || Number.isNaN(val))) {
+    if (propSchema.type === "number" && (typeof val !== "number" || !Number.isFinite(val))) {
       return { valid: false, error: `Parameter '${key}' must be a number.` };
     }
+    if (key === "locale" && typeof val === "string") {
+      if (!normalizeLocaleTag(val)) {
+        return {
+          valid: false,
+          error: `Parameter 'locale' must match BCP-47 for one of: ${SUPPORTED_LOCALE_LIST.join(", ")}.`,
+        };
+      }
+      continue;
+    }
+    if (typeof propSchema.pattern === "string" && typeof val === "string") {
+      const regex = new RegExp(propSchema.pattern);
+      if (!regex.test(val)) {
+        return { valid: false, error: `Parameter '${key}' does not match required pattern.` };
+      }
+    }
     if (Array.isArray(propSchema.enum) && !propSchema.enum.includes(val)) {
       return {
         valid: false,
@@ -550,6 +631,87 @@ export function validateToolArguments(toolName, args = {}) {
     }
   }
 
+  if (toolName === "systemores_export_dossier") {
+    if (args.targetFrameworks !== undefined) {
+      const validFrameworks = new Set(["eu_ai_act", "nist_ai_rmf", "eeoc_ecoa", "soc2_ai"]);
+      const tokens = args.targetFrameworks
+        .split(",")
+        .map((s) => s.trim())
+        .filter(Boolean);
+      if (tokens.length === 0 || tokens.some((fw) => !validFrameworks.has(fw))) {
+        return {
+          valid: false,
+          error:
+            "Parameter 'targetFrameworks' must contain one or more of: eu_ai_act, nist_ai_rmf, eeoc_ecoa, soc2_ai.",
+        };
+      }
+    }
+
+    if (args.systemPrompt !== undefined) {
+      if (!args.systemPrompt.trim() || args.systemPrompt.length > 8000) {
+        return {
+          valid: false,
+          error: "Parameter 'systemPrompt' must be a non-empty string up to 8,000 characters.",
+        };
+      }
+    }
+
+    for (const jsonField of ["toolDefinitionsJson", "graphTopologyJson", "otelTraceJson"]) {
+      if (args[jsonField] !== undefined && (!args[jsonField].trim() || args[jsonField].length > 100000)) {
+        return {
+          valid: false,
+          error: `Parameter '${jsonField}' must be a non-empty JSON string up to 100,000 characters.`,
+        };
+      }
+    }
+
+    if (args.liveAuditMetricsJson !== undefined) {
+      if (!args.liveAuditMetricsJson.trim() || args.liveAuditMetricsJson.length > 10000) {
+        return {
+          valid: false,
+          error: "Parameter 'liveAuditMetricsJson' must be a non-empty JSON string up to 10,000 characters.",
+        };
+      }
+      let parsedMetrics;
+      try {
+        parsedMetrics = JSON.parse(args.liveAuditMetricsJson);
+      } catch {
+        return {
+          valid: false,
+          error: "Parameter 'liveAuditMetricsJson' must be valid JSON.",
+        };
+      }
+      if (!parsedMetrics || typeof parsedMetrics !== "object" || Array.isArray(parsedMetrics)) {
+        return {
+          valid: false,
+          error: "Parameter 'liveAuditMetricsJson' must be a JSON object.",
+        };
+      }
+      const bounds = {
+        healthIndex: [0, 100],
+        sycophancyProbability: [0, 1],
+        personaStabilityIndex: [0, 1],
+        averageDriftIndex: [0, 1],
+        toolIntegrityScore: [0, 100],
+      };
+      for (const [metricKey, [min, max]] of Object.entries(bounds)) {
+        if (!Object.hasOwn(parsedMetrics, metricKey) || parsedMetrics[metricKey] === undefined) continue;
+        const metricVal = parsedMetrics[metricKey];
+        if (
+          typeof metricVal !== "number" ||
+          !Number.isFinite(metricVal) ||
+          metricVal < min ||
+          metricVal > max
+        ) {
+          return {
+            valid: false,
+            error: `Parameter 'liveAuditMetricsJson.${metricKey}' must be a finite number between ${min} and ${max}.`,
+          };
+        }
+      }
+    }
+  }
+
   return { valid: true };
 }
 
@@ -587,6 +749,12 @@ export async function executeToolCall(toolName, args = {}, options = {}) {
     };
   }
 
+  const forwardArgs =
+    args && typeof args === "object" && !Array.isArray(args) ? { ...args } : {};
+  if (typeof forwardArgs.locale === "string") {
+    forwardArgs.locale = normalizeLocaleTag(forwardArgs.locale) || "en";
+  }
+
   if (options.dryRun) {
     return {
       isError: false,
@@ -612,30 +780,60 @@ export async function executeToolCall(toolName, args = {}, options = {}) {
     headers.Authorization = `Bearer ${auth.token}`;
   }
 
-  const response = await fetch(REMOTE_MCP_ENDPOINT, {
-    method: "POST",
-    headers,
-    body: JSON.stringify({
-      jsonrpc: "2.0",
-      id: Date.now(),
-      method: "tools/call",
-      params: { name: toolName, arguments: args },
-    }),
-  });
-
-  const payload = await response.json();
-  if (payload.result) {
-    return payload.result;
+  try {
+    const response = await fetch(REMOTE_MCP_ENDPOINT, {
+      method: "POST",
+      signal: AbortSignal.timeout(15_000),
+      headers,
+      body: JSON.stringify({
+        jsonrpc: "2.0",
+        id: Date.now(),
+        method: "tools/call",
+        params: { name: toolName, arguments: forwardArgs },
+      }),
+    });
+
+    const rawText = await response.text();
+    let payload = {};
+    if (rawText.trimStart().startsWith("{")) {
+      payload = JSON.parse(rawText);
+    } else {
+      const sseDataLines = rawText
+        .split("\n")
+        .filter((line) => line.startsWith("data: "))
+        .map((line) => line.slice(6).trim())
+        .filter((chunk) => chunk && chunk !== "[DONE]");
+      if (sseDataLines.length > 0) {
+        payload = JSON.parse(sseDataLines[sseDataLines.length - 1]);
+      }
+    }
+    if (payload.result) {
+      return payload.result;
+    }
+    const errText =
+      typeof payload.error === "string"
+        ? payload.error
+        : payload.error?.message || `Remote MCP error (HTTP ${response.status})`;
+    return {
+      isError: true,
+      content: [
+        {
+          type: "text",
+          text: errText,
+        },
+      ],
+    };
+  } catch (error) {
+    return {
+      isError: true,
+      content: [
+        {
+          type: "text",
+          text: `Remote MCP transport error: ${error instanceof Error ? error.message : String(error)}`,
+        },
+      ],
+    };
   }
-  return {
-    isError: true,
-    content: [
-      {
-        type: "text",
-        text: payload.error?.message || `Remote MCP error (HTTP ${response.status})`,
-      },
-    ],
-  };
 }
 
 // ─── MCP SDK Static AST & Runtime Registration Bridge ────────────────────────
@@ -644,7 +842,7 @@ export const CallToolRequestSchema = { method: "tools/call" };
 
 export const server = {
   name: "systemores-plugin",
-  version: "1.1.3",
+  version: "1.2.0",
   registeredTools: new Map(),
   handlers: new Map(),
   tool(name, description, schema, annotations, handler) {
@@ -695,7 +893,7 @@ export async function handleJsonRpcMessage(message) {
         },
         serverInfo: {
           name: "systemores-plugin",
-          version: "1.1.3",
+          version: "1.2.0",
         },
       },
     };
@@ -764,6 +962,17 @@ export function startStdioServer() {
   });
 }
 
-if (process.argv[1] && fileURLToPath(import.meta.url) === process.argv[1]) {
+export function isExecutedDirectly(metaUrl, argv1 = process.argv[1]) {
+  if (!argv1) return false;
+  const modulePath = fileURLToPath(metaUrl);
+  if (modulePath === argv1) return true;
+  try {
+    return realpathSync(modulePath) === realpathSync(argv1);
+  } catch {
+    return false;
+  }
+}
+
+if (isExecutedDirectly(import.meta.url)) {
   startStdioServer();
 }