@oomkapwn/enquire-mcp 3.6.2 → 3.6.3

package/CHANGELOG.md

@@ -2,6 +2,44 @@
 
 All notable changes to this project will be documented here. The format follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and the project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [3.6.3] — 2026-05-15
+
+> **TL;DR:** Discoverability + positioning patch. README, npm description, `package.json#keywords`, and the GitHub repo About + Topics now lead with **"long-term memory for AI agents"** framing — aligning with the post-Claude-Memory (Oct 2025) / post-Anthropic-Skills (Nov 2025) developer-discovery vocabulary. The capability we've shipped since v1.0 (durable, queryable, vendor-neutral memory in plain markdown) hasn't changed — just the framing on the discovery surfaces. **No code, API, schema, or behavior changes. 753 tests still passing.** No-op upgrade for everyone.
+
+**Patch — discoverability / SEO repositioning (zero functional changes).**
+
+### Changed — discovery & positioning
+
+- **README headline subtitle**: `"Every modern IR primitive. In one tool. For free."` → `"The most advanced Obsidian MCP — every modern IR primitive, in one tool, for free."` (preserved as secondary line). New PRIMARY headline: **"Long-term memory for AI agents. Built on your Obsidian vault."**
+- **README `## What it is` lead paragraph**: rewritten to frame the project as the **open-source, MCP-native, agent-grade memory layer** that complements Claude Memory / ChatGPT Memory / Cursor memory with vendor-neutral, file-owned, MCP-portable persistence. All technical claims preserved + sourced.
+- **README new `## 🧠 Use cases` section** (between Quick start and API reference). 3 explicit scenarios — (1) long-term memory for AI agents, (2) personal knowledge base / second brain, (3) agentic RAG / context engineering — each anchored to specific enquire-mcp capabilities (hybrid retrieval, multilingual, graph-boost, HyDE, sub-question decomposition, eval harness).
+- **`package.json#description`**: now leads with `"Memory layer for AI agents over your Obsidian vault."`. All capability claims preserved. Adds the phrase "open-source long-term memory / second brain".
+- **`package.json#keywords`** (60 → 71): added 10 hype-aligned keywords at the **top of the array** (npm search ranks early keywords higher in many surfaces): `ai-memory`, `agent-memory`, `llm-memory`, `long-term-memory`, `claude-memory`, `memory-for-ai-agents`, `context-engineering`, `second-brain`, `personal-knowledge-base`, `knowledge-graph`. Also added `ai-agents` (plural) and `hybrid-search` (was only `hybrid-retrieval`). All 60 prior keywords retained.
+- **GitHub repo About** (via `gh api -X PATCH /repos/oomkapwn/enquire-mcp`): updated to `"Memory layer for AI agents over your Obsidian vault. Hybrid retrieval (BM25 + ML + BGE rerank, RRF-fused), HNSW + int8 quantization, agentic RAG (HyDE + sub-question), standalone Bases, PDFs+OCR. Open-source long-term memory for Claude Code/Desktop, Cursor, ChatGPT, Codex. MCP-native, MIT, SLSA-3."`
+- **GitHub Topics** (rebalanced within the 20-cap): added 7 hype topics — `ai-memory`, `agent-memory`, `llm-memory`, `long-term-memory`, `claude-memory`, `second-brain`, `context-engineering`. Dropped 7 lower-yield existing topics (`hnsw`, `ocr`, `pdf`, `openclaw`, `embeddings`, `vector-search`, `claude-desktop` — all still discoverable via npm keywords + README). Final 20-topic set: `obsidian, obsidian-mcp, mcp-server, model-context-protocol, claude, claude-code, cursor, chatgpt, codex, rag, agentic-rag, hybrid-search, semantic-search, ai-memory, agent-memory, llm-memory, long-term-memory, claude-memory, second-brain, context-engineering`.
+
+### Tests
+
+**753 tests** — identical to v3.6.2. No code paths changed → no test additions / removals / regressions. Lint clean, `tsc` strict + `noUncheckedIndexedAccess` clean, version-consistency green at `3.6.3` (5 surfaces), changelog-coverage gate passes (this section makes no coverage claims, so nothing to check).
+
+### Migration
+
+**No-op for every consumer.** Identical public API (44 tools), CLI, `package.json#exports`, dependency tree, on-disk DB schema, MCP wire format. Existing README anchors and links are preserved; the new `## 🧠 Use cases` section is **additive** between Quick start and API reference.
+
+### Method note
+
+This patch is a **deliberate timing call**, not a fix for a found defect.
+
+In Oct 2025 Anthropic shipped **Claude Memory**; in Nov 2025 **Claude Skills**. Both moved phrases like *"memory for AI agents"*, *"long-term agent memory"*, *"context engineering"* from niche jargon into mainstream developer-discovery vocabulary on npm, GitHub, X, and Google. enquire-mcp has shipped the underlying primitive — durable, queryable, vendor-neutral long-term memory in plain markdown — since v1.0 (public OSS release 2026-05-02; my private vault dogfooding goes back to early 2025). The technical capability hasn't changed; the search demand around the *words for that capability* has.
+
+The lesson: **positioning isn't a one-time launch decision — it's continuous calibration against where the audience's vocabulary actually IS.** v3.6.x already had the strongest IR stack in the open-source Obsidian-MCP space (per `docs/benchmarks.md` + `docs/COMPARISON.md`). This patch makes that fact findable by the people actively searching for *"agent memory"* / *"long-term memory for Claude"* / *"second brain for Cursor"*.
+
+**Non-goal: no over-claiming, no spec inflation.** Every capability claim in the new framing maps to an empirical benchmark in `docs/benchmarks.md` or a row in the comparison matrix. The memory-positioning is layered *on top of* the technical claims, never replacing them. Success criterion: ≥1pp lift in repo / npm impression CTR over the 14 days following 2026-05-15 (will verify via npm search analytics + GitHub Insights traffic; if no measurable lift by 2026-05-29 the framing returns to evaluation).
+
+**Why a patch release and not just a README edit on `main`:** the `package.json#description` and `package.json#keywords` changes are surfaces visible on **npmjs.com** — and those only update when a new version publishes. So the marketing pivot must ship as a real version bump to land on npm. Going with `patch` (not minor) because the public API surface and behavior are byte-identical to v3.6.2.
+
+---
+
 ## [3.6.2] — 2026-05-15
 
 > **TL;DR:** Audit batch — closes **K-1 RESIDUAL CLASS** (the v3.6.1 CRIT-1 fix was instance-only; the destructive-bootstrap-schema class was still active in 4 hot paths including a sibling K-1b in FtsIndex tokenize_mode). Plus 13 Medium + 14 Low findings from the internal 9-layer audit + 4 HIGHs from a second external audit. **+37 tests** (753 total, +1.29pp branches margin). No breaking API changes. **Retroactive correction** — v3.6.1's "CRIT-1 closed" was an overclaim.

package/README.md

@@ -4,9 +4,9 @@
 
 # enquire-mcp
 
-### The most advanced Obsidian MCP server. Period.
+### Long-term memory for AI agents. Built on your Obsidian vault.
 
-**Every modern IR primitive. In one tool. For free.**
+**The most advanced Obsidian MCP — every modern IR primitive, in one tool, for free.**
 
 [![CI](https://github.com/oomkapwn/enquire-mcp/actions/workflows/ci.yml/badge.svg)](https://github.com/oomkapwn/enquire-mcp/actions/workflows/ci.yml)
 [![npm](https://img.shields.io/npm/v/@oomkapwn/enquire-mcp.svg?label=npm&color=cb3837)](https://www.npmjs.com/package/@oomkapwn/enquire-mcp)
@@ -27,7 +27,9 @@
 
 ## What it is
 
-A **production-ready MCP server** that gives any AI agent — Claude Code, Claude Desktop, Cursor, ChatGPT custom GPT, Codex, mobile MCP clients — structured access to your Obsidian vault. The umbrella `obsidian_search` tool fuses **BM25 + TF-IDF + multilingual ML embeddings** via Reciprocal Rank Fusion (Cormack et al, 2009), reranks with a **BGE cross-encoder** (5 model options), scales to millions of chunks via **HNSW with int8 quantization**, and returns blended markdown + PDF hits with `[page: N]` citations.
+A **production-ready MCP server** that turns your Obsidian vault into **persistent, queryable long-term memory** for any AI agent — Claude Code, Claude Desktop, Cursor, ChatGPT custom GPT, Codex, mobile MCP clients. Unlike session-scoped chat memory or proprietary cloud "memory" features, your knowledge lives in plain markdown files **you own**, indexed locally, and recalled with the full modern IR stack: **BM25 + TF-IDF + multilingual ML embeddings** fused via Reciprocal Rank Fusion (Cormack et al, 2009), reranked with a **BGE cross-encoder** (5 model options), scaled to millions of chunks via **HNSW with int8 quantization**, with blended markdown + PDF hits and `[page: N]` citations.
+
+Think of it as the **open-source, MCP-native, agent-grade memory layer** that complements Claude Memory / ChatGPT Memory / Cursor memory — but stores your durable knowledge in a portable, vendor-neutral format (your Obsidian vault) any agent can read.
 
 **44 tools · 19 MCP prompts · 753 unit tests · 50+ languages · v3.6.x · semver-bound · MIT · SLSA-3.**
 
@@ -65,6 +67,16 @@ enquire-mcp doctor --vault <path>    # color-coded ✓/⚠/✗ health check
 
 ---
 
+## 🧠 Use cases
+
+**1 — Long-term memory for AI agents.** Drop your Obsidian vault into any MCP-compatible agent (Claude Code, Claude Desktop, Cursor, ChatGPT, Codex). The agent now has durable, semantic recall over every meeting note, journal entry, research log, and decision doc you've ever written — across sessions, models, and providers. Unlike `Claude Memory` or `ChatGPT Memory`, your knowledge isn't locked into one vendor's cloud; it lives in plain markdown you own and can migrate freely.
+
+**2 — Personal knowledge base / second brain.** Hybrid retrieval surfaces the right note for *any* phrasing, in any of 50+ languages. Ask in English about a Russian-language journal entry from 2 years ago, get the right hit. Wikilink graph-boost reranks notes that sit at the centre of your knowledge graph. GraphRAG-light surfaces topical communities — discover connections you forgot you made. PDFs blend into search with `[page: N]` citations so research papers and meeting transcripts become first-class memory.
+
+**3 — Agentic RAG / context engineering.** `obsidian_search` exposes per-signal scores so the agent sees *why* each hit ranked. HyDE pre-rewrites vague queries into rich hypothetical answers before retrieval. Sub-question decomposition handles multi-hop questions ("how did our pricing strategy evolve and what was the customer reaction?") by breaking them into independent sub-queries, fusing results. The built-in eval harness (NDCG / Recall / MRR) lets you measure retrieval quality on your own queries instead of trusting vendor benchmarks.
+
+---
+
 ## 📖 API reference
 
 Auto-generated **[API reference at oomkapwn.github.io/enquire-mcp](https://oomkapwn.github.io/enquire-mcp/)** — every tool, prompt, and exported helper with full TSDoc (`@param` / `@returns` / `@example`). Rebuilt from source on every push to `main` via [`publish-docs.yml`](./.github/workflows/publish-docs.yml) (TypeDoc → GitHub Pages). Drift-free by construction: the same TSDoc that AI agents and IDEs see is what's published.

package/dist/index.d.ts

@@ -7,7 +7,7 @@
  * + `McpServer({version})`) and `src/tool-registry.ts` (used in the
  * `vault-info` resource payload).
  */
-export declare const VERSION = "3.6.2";
+export declare const VERSION = "3.6.3";
 export { main } from "./cli.js";
 export { buildEmbedText, buildMcpServer, formatReadyBanner, prepareServerDeps, type ServeOptions, type ServerDeps, startServer } from "./server.js";
 export { parsePositiveInt, parseQuantizationMode } from "./tool-registry.js";

package/dist/index.js

@@ -32,7 +32,7 @@ import { main } from "./cli.js";
  * + `McpServer({version})`) and `src/tool-registry.ts` (used in the
  * `vault-info` resource payload).
  */
-export const VERSION = "3.6.2";
+export const VERSION = "3.6.3";
 // Re-exports — preserve the v3.5.x public surface so http-transport.ts and
 // tests don't need to know about the new module layout. The set below
 // exactly matches the v3.5.x `export` declarations: `main`,