prism-mcp-server 19.2.2 → 19.2.4

package/README.md

@@ -93,7 +93,7 @@ Every `prism_infer` call tracks which model handled it (local Ollama vs cloud) a
     synalux-27b: 2 calls, 1,500 tokens, avg 1,100ms
 ```
 
-Local calls use actual Ollama token counts; cloud calls use estimates. Metrics are aggregated by the Synalux portal — Prism is a thin client that forwards per-call data and fetches the summary on demand.
+Local calls use actual Ollama token counts (`prompt_eval_count` / `eval_count` from Ollama); cloud calls use char/4 estimates. Metrics are tracked locally — no portal dependency, no env vars, works offline. Per-call data is also forwarded to the Synalux portal as best-effort analytics (independent of the display).
 
 ### Session Drift Detection
 
@@ -363,7 +363,7 @@ All on-device models are free to run locally via Ollama on every tier. A subscri
 | Cloud Coder (Web IDE) | -- | 100/day | 1,000/day | 100,000/day |
 | Cloud search | -- | 50/day | 500/day | 100,000/day |
 | Max output tokens | 512 | 1,024 | 2,048 | 4,096 |
-| Cloud fallback | -- | Claude Sonnet 4 | Claude Sonnet 4 | Priority + Sonnet 4 |
+| Cloud fallback | -- | Claude Opus 4.7 | Claude Opus 4.7 | Priority + Opus 4.7 |
 | Grounding verifier (fact-check AI output) | -- | ✅ | ✅ | ✅ |
 | Memory sync (cloud) | -- | ✅ | ✅ | ✅ |
 | Knowledge / session memory | limited | unlimited | unlimited | unlimited |
@@ -389,6 +389,7 @@ Prism exposes 40+ MCP tools. The core memory loop:
 | `verify_behavior` | Pre-edit scenario challenge — catch bad changes before they happen |
 | `knowledge_ingest` | Teach Prism a codebase or document |
 | `prism_infer` | Local-first inference (route/chat/code modes, thinking, cloud escalation) |
+| `inference_metrics` | Session delegation stats on demand (call count, tokens, local/cloud split) |
 
 ### `prism_infer` — local-first inference with cloud escalation
 
@@ -553,6 +554,7 @@ Routing is automatic: `9b → 4b → cloud fallback` on desktop/server, `2b →
 | `PRISM_SYNALUX_API_KEY` | Paid-tier portal key (`synalux_sk_...`) | -- (local if unset) |
 | `LOCAL_LLM_URL` | Ollama endpoint | `http://localhost:11434` |
 | `PRISM_FORCE_LOCAL` | Force local SQLite regardless of credentials | `false` |
+| `TELEMETRY_WRITE_TOKEN` | Portal analytics token (optional — metrics display works without it) | -- |
 
 With no variables set, Prism runs fully local. Set `PRISM_SYNALUX_API_KEY` (and leave `PRISM_STORAGE=auto`) to use the cloud backend.
 
@@ -561,11 +563,11 @@ With no variables set, Prism runs fully local. Set `PRISM_SYNALUX_API_KEY` (and
 ## Testing
 
 ```bash
-npm test                 # full suite (vitest)
+npm test                 # full suite (vitest) — 95 files, 2841 tests
 npm test -- --coverage   # coverage report
 ```
 
-Coverage spans HRR retrieval, knowledge ingestion, the inference cascade and grounding verifier, compaction, the model picker, and storage round-trips.
+Coverage spans HRR retrieval, knowledge ingestion, the inference cascade and grounding verifier, inference metrics, telemetry allowlist, delegation gate, compaction, the model picker, and storage round-trips.
 
 ---
 

package/dist/server.js

@@ -79,6 +79,7 @@ import { sanitizeMcpOutput } from "./utils/sanitizer.js";
 import { getTracer, initTelemetry } from "./utils/telemetry.js";
 import { context as otelContext, trace, SpanStatusCode } from "@opentelemetry/api";
 import { ddInfo, ddError as ddLogError } from "./utils/ddLogger.js";
+import { inferenceMetricsHandler } from "./utils/inferenceMetrics.js";
 // ─── Import Tool Definitions (schemas) and Handlers (implementations) ─────
 import { WEB_SEARCH_TOOL, BRAVE_WEB_SEARCH_CODE_MODE_TOOL, LOCAL_SEARCH_TOOL, BRAVE_LOCAL_SEARCH_CODE_MODE_TOOL, CODE_MODE_TRANSFORM_TOOL, BRAVE_ANSWERS_TOOL, RESEARCH_PAPER_ANALYSIS_TOOL, webSearchHandler, braveWebSearchCodeModeHandler, localSearchHandler, braveLocalSearchCodeModeHandler, codeModeTransformHandler, braveAnswersHandler, researchPaperAnalysisHandler, } from "./tools/index.js";
 // Session memory tools — only used if Supabase is configured
@@ -112,7 +113,9 @@ VERIFY_BEHAVIOR_TOOL, isVerifyBehaviorArgs,
 // v12: Developer Onboarding & Enterprise Observability
 ONBOARDING_WIZARD_TOOL, EXTRACT_ENTITIES_TOOL, API_ANALYTICS_TOOL, BACKUP_DATABASE_TOOL, CONFIGURE_NOTIFICATIONS_TOOL, QUERY_MEMORY_NATURAL_TOOL, 
 // v15.5: Knowledge Ingestion
-KNOWLEDGE_INGEST_TOOL, sessionSaveLedgerHandler, sessionSaveHandoffHandler, sessionLoadContextHandler, knowledgeSearchHandler, knowledgeForgetHandler, 
+KNOWLEDGE_INGEST_TOOL, 
+// v19.2: Inference Metrics
+INFERENCE_METRICS_TOOL, sessionSaveLedgerHandler, sessionSaveHandoffHandler, sessionLoadContextHandler, knowledgeSearchHandler, knowledgeForgetHandler, 
 // ─── v0.4.0: New tool handlers ───
 compactLedgerHandler, sessionSearchMemoryHandler, backfillEmbeddingsHandler, sessionBackfillLinksHandler, sessionSynthesizeEdgesHandler, sessionCognitiveRouteHandler, 
 // ─── v2.0: Time Travel handlers ───
@@ -246,6 +249,8 @@ function buildSessionMemoryTools(autoloadList) {
         QUERY_MEMORY_NATURAL_TOOL, // query_memory_natural — NL → structured memory search
         // ─── v15.5: Knowledge Ingestion ───
         KNOWLEDGE_INGEST_TOOL, // knowledge_ingest — chunk code, gen Q&A, store in graph
+        // ─── v19.2: Inference Metrics ───
+        INFERENCE_METRICS_TOOL, // inference_metrics — read-only session delegation stats
     ];
 }
 // ─── v0.4.0: Resource Subscription Tracking ──────────────────────
@@ -960,6 +965,9 @@ export function createServer() {
                             throw new Error("Session memory not configured.");
                         result = await knowledgeIngestHandler(args);
                         break;
+                    case "inference_metrics":
+                        result = await inferenceMetricsHandler();
+                        break;
                     default:
                         result = {
                             content: [{ type: "text", text: `Unknown tool: ${name}` }],

package/dist/tools/index.js

@@ -63,6 +63,7 @@ export { verifyBehaviorHandler } from "./behavioralVerifierHandler.js";
 // Chunks source code, generates Q&A via Claude Haiku, stores in knowledge graph.
 // Three entry points: MCP tool, REST API, GitHub webhook.
 export { KNOWLEDGE_INGEST_TOOL } from "./ingestDefinitions.js";
+export { INFERENCE_METRICS_TOOL } from "./sessionMemoryDefinitions.js";
 export { knowledgeIngestHandler, handleGitHubWebhook, ingestKnowledge, isIngestArgs } from "./ingestHandler.js";
 // ── v15.4: prism_infer — local-first inference (RAM-gated cascade) ──
 // Always available. Saves caller's cloud tokens by routing to local

package/dist/tools/sessionMemoryDefinitions.js

@@ -1790,3 +1790,15 @@ export function isVerifyBehaviorArgs(a) {
         return false;
     return true;
 }
+// ─── v19.2: Inference Metrics Tool ──────────────────────────
+export const INFERENCE_METRICS_TOOL = {
+    name: "inference_metrics",
+    description: "Returns the current session's local-model inference metrics — call count, " +
+        "local vs cloud split, token totals, per-model breakdown, and average latency. " +
+        "Read-only, no arguments. Reflects prism_infer delegation usage only, not the " +
+        "host model's (Claude's) own token spend (use /cost for that).",
+    inputSchema: {
+        type: "object",
+        properties: {},
+    },
+};

package/dist/utils/inferenceMetrics.js

@@ -70,12 +70,21 @@ export function resetInferenceMetrics() {
     }
     debugLog("[inference-metrics] Session metrics reset");
 }
+export async function inferenceMetricsHandler() {
+    const block = formatInferenceMetrics();
+    return {
+        content: [{
+                type: "text",
+                text: block || "No prism_infer calls this session. Metrics track local-model delegation only — not the host model's (Claude's) token spend.",
+            }],
+    };
+}
 export function formatInferenceMetrics() {
     const snap = getInferenceSnapshot();
     if (snap.totalCalls === 0)
         return "";
     const lines = [
-        `\n📊 Inference Metrics (this session):`,
+        `\n📊 Inference Metrics — local-model delegation (this session):`,
         `  Total calls: ${snap.totalCalls} — Local: ${snap.localCalls} (${snap.localPct}%) | Cloud: ${snap.cloudCalls} (${snap.cloudPct}%)`,
         `  Tokens: ${snap.totalPromptTokens.toLocaleString()} in + ${snap.totalCompletionTokens.toLocaleString()} out = ${snap.totalTokens.toLocaleString()} total`,
         `  Avg latency: ${snap.avgLatencyMs}ms`,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "prism-mcp-server",
-  "version": "19.2.2",
+  "version": "19.2.4",
   "mcpName": "io.github.dcostenco/prism-coder",
   "description": "Prism Coder — Cognitive memory + tool-calling intelligence for AI agents. Mind Palace persistent memory (BFCL Gold Certified, 100% Tool-Call Accuracy, 114 Agent Skills, PHI Guard, Tier Enforcement, Prompt-Based Skill Routing, Zero-Search HDC/HRR retrieval, HRR Semantic Drift Detection across BCBA/Coding/AAC domains, HIPAA-hardened local-first storage, SLERP-optimized GRPO alignment) plus the prism-coder 1.7B–32B open-weights LLM fleet.",
   "module": "index.ts",