mcp-event-intelligence 0.3.0 → 0.3.1

package/CHANGELOG.md

@@ -2,6 +2,24 @@
 
 All notable public changes are documented here.
 
+## [0.3.1] - 2026-09-20
+
+Public-polish and reliability patch before wider distribution.
+
+### Reliability
+
+- fixed the lease-claim test cleanup race that could surface intermittent `ENOTEMPTY` errors while an in-flight worker was still writing;
+- added repeated direct lease-claim contention coverage and full coordinator contention stress coverage.
+
+### Documentation and public surface
+
+- refreshed Security Policy and Security Model for scoped multi-tenant hosts, Activation Envelopes and the v0.3 agent-friendly planning flow;
+- clarified the distinction between package version 0.3.x and the intentionally compatible internal protocol/schema v0.1 identifiers;
+- refreshed Architecture and Protocol docs for `planTrigger()`, continuation contracts and wake hydration;
+- updated MCP Registry docs to reflect the package and Registry as live, not deferred;
+- updated the CLI help banner;
+- removed the obsolete one-shot v0.1.0 release workflow.
+
 ## [0.3.0] - 2026-09-20
 
 Agent-integration release focused on making Event Intelligence usable as a plug-and-play future-condition primitive.

package/README.md

@@ -4,7 +4,7 @@
 
 [![CI](https://github.com/sarooo17/event-intelligence/actions/workflows/ci.yml/badge.svg)](https://github.com/sarooo17/event-intelligence/actions/workflows/ci.yml)
 [![npm](https://img.shields.io/npm/v/mcp-event-intelligence.svg)](https://www.npmjs.com/package/mcp-event-intelligence)
-[![MCP Registry](https://img.shields.io/badge/MCP%20Registry-v0.3.0-5b5bd6)](https://registry.modelcontextprotocol.io/?q=io.github.sarooo17%2Fevent-intelligence)
+[![MCP Registry](https://img.shields.io/badge/MCP%20Registry-v0.3.1-5b5bd6)](https://registry.modelcontextprotocol.io/?q=io.github.sarooo17%2Fevent-intelligence)
 [![License](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](LICENSE)
 
 MCP Event Intelligence is an experimental event runtime for agents that need to react to **future conditions over multiple event sources** without keeping an LLM or agent loop alive.
@@ -300,7 +300,7 @@ Event Intelligence has one optional AI boundary: **semantic correlation**.
 - embedded hosts may inject a compatible `semanticEvaluator` instead.
 - there is no bundled OpenAI planner and no OpenAI API dependency.
 
-The agent/harness is already responsible for reasoning and can author a structured trigger directly from `event_sources_list` / the discovered schemas. Deterministic correlation, temporal logic, persistence, derived events and wake delivery require no model API.
+The agent/harness is already responsible for natural-language reasoning. It can use `trigger_plan` / `planTrigger()` to compile common requests deterministically against discovered source schemas, or submit a canonical trigger definition directly for advanced cases. Deterministic correlation, temporal logic, persistence, derived events and wake delivery require no model API.
 
 ## Full-system acceptance
 
@@ -350,13 +350,13 @@ The standalone service supports manual/provider-native event ingress. It does no
 ### Docker
 
 ```bash
-docker build -t mcp-event-intelligence:0.3.0 .
+docker build -t mcp-event-intelligence:0.3.1 .
 
 docker run --rm \
   -p 3000:3000 \
   -v mcp-event-intelligence-data:/data \
   -e SERVICE_AUTH_TOKEN="$(openssl rand -hex 32)" \
-  mcp-event-intelligence:0.3.0
+  mcp-event-intelligence:0.3.1
 ```
 
 ## Optional MCP control plane
@@ -445,7 +445,7 @@ io.github.sarooo17/event-intelligence
 
 ## Project status
 
-**v0.3 reference implementation / experimental.**
+**v0.3.x reference implementation / experimental.**
 
 The architecture is implemented and exercised end-to-end. Storage is now injectable and scoped, while the bundled JSONL backend remains a single-process reference implementation. Remaining work is primarily production database adapters/HA validation, scale benchmarks and upstream feedback.
 

package/SECURITY.md

@@ -2,12 +2,12 @@
 
 ## Supported versions
 
-The public reference implementation is currently pre-1.0.
+The public package is pre-1.0. Security fixes are supported on the current 0.3.x line.
 
 | Version | Supported |
 | --- | --- |
-| 0.1.x | Yes |
-| < 0.1 | No |
+| 0.3.x | Yes |
+| < 0.3 | No |
 
 ## Reporting a vulnerability
 
@@ -26,17 +26,21 @@ Useful reports include:
 
 ## Security boundaries
 
-Event Intelligence v0.1 is a reference implementation, not a hardened multi-tenant hosted service.
+Event Intelligence 0.3.x is an experimental reference implementation with scoped multi-tenant host support. It is not a turnkey hardened hosted service.
 
 Important defaults and constraints:
 
-- protected control-plane endpoints require `SERVICE_AUTH_TOKEN`;
+- tenant/workspace isolation is implemented through scoped store partitions; tenant-facing code should receive only the corresponding scoped host view;
+- protected standalone control-plane endpoints require `SERVICE_AUTH_TOKEN`;
 - GitHub ingress verifies provider HMAC when configured;
-- runtime wake packets are HMAC-signed and runtime targets are operator-configured;
+- standalone runtime wake packets are HMAC-signed and runtime targets are operator-configured;
+- embedded wakes preserve the configured continuation separately from untrusted matched-event evidence;
+- event receipt never grants new downstream authorization;
 - arbitrary public HTTP wake targets are rejected;
 - derived-event lineage is refs-only by default;
 - audit records are hash-linked and tamper-evident, not tamper-proof;
-- the JSONL store is single-writer and not safe for horizontal multi-replica deployment;
+- the bundled JSONL store is single-process reference infrastructure and is not safe as a shared multi-replica backend;
+- horizontally scaled custom stores must preserve scope isolation and implement wake claim/lease operations atomically;
 - secrets belong in environment variables / secret managers, never committed files.
 
 See `docs/SECURITY-MODEL.md` for the architectural model.

package/bin/mcp-event-intelligence.mjs

@@ -4,12 +4,12 @@ const args = process.argv.slice(2);
 const flags = new Set(args);
 
 if (flags.has('--version') || flags.has('-v')) {
-  console.log('0.3.0');
+  console.log('0.3.1');
   process.exit(0);
 }
 
 if (flags.has('--help') || flags.has('-h')) {
-  console.log(`MCP Event Intelligence v0.1
+  console.log(`MCP Event Intelligence v0.3
 
 Durable temporal event intelligence for sleeping agents.
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mcp-event-intelligence",
-  "version": "0.3.0",
+  "version": "0.3.1",
   "description": "Durable temporal event intelligence for agent hosts: host-owned MCP event discovery, composite triggers, derived events, and targeted wakeups.",
   "license": "Apache-2.0",
   "type": "module",

package/scripts/mcp-stdio-server.mjs

@@ -451,7 +451,7 @@ export async function buildEventIntelligenceMcpServer({
   const server = new McpServer(
     {
       name: 'mcp-event-intelligence',
-      version: '0.3.0',
+      version: '0.3.1',
     },
     {
       instructions:

package/server.json

@@ -3,7 +3,7 @@
   "name": "io.github.sarooo17/event-intelligence",
   "title": "MCP Event Intelligence",
   "description": "Durable temporal event intelligence: composite triggers, derived events, and runtime wakeups.",
-  "version": "0.3.0",
+  "version": "0.3.1",
   "repository": {
     "url": "https://github.com/sarooo17/event-intelligence",
     "source": "github"
@@ -13,7 +13,7 @@
       "registryType": "npm",
       "registryBaseUrl": "https://registry.npmjs.org",
       "identifier": "mcp-event-intelligence",
-      "version": "0.3.0",
+      "version": "0.3.1",
       "transport": {
         "type": "stdio"
       },