@gmickel/gno 0.10.4 → 0.12.0

package/README.md

@@ -27,6 +27,16 @@ GNO is a local knowledge engine for privacy-conscious developers and AI agents.
 
 ---
 
+## What's New in v0.12
+
+- **Note Linking**: Wiki-style `[[links]]`, backlinks, and AI-powered related notes
+- **Tag System**: Filter searches by frontmatter tags with `--tags-any`/`--tags-all`
+- **Web UI**: Outgoing links panel, backlinks panel, related notes sidebar
+- **CLI**: `gno links`, `gno backlinks`, `gno similar` commands
+- **MCP**: New tools for AI agents to navigate your knowledge graph
+
+---
+
 ## Quick Start
 
 ```bash
@@ -315,6 +325,7 @@ graph TD
 | **MCP Server**      | Works with Claude Desktop, Cursor, Zed, + 8 more                               |
 | **Collections**     | Organize sources with patterns, excludes, contexts                             |
 | **Tag Filtering**   | Frontmatter tags with hierarchical paths, filter via `--tags-any`/`--tags-all` |
+| **Note Linking**    | Wiki links, backlinks, related notes, cross-collection navigation              |
 | **Multilingual**    | 30+ languages, auto-detection, cross-lingual search                            |
 | **Incremental**     | SHA-256 tracking, only changed files re-indexed                                |
 | **Keyboard First**  | ⌘N capture, ⌘K search, ⌘/ shortcuts, ⌘S save                                   |

package/assets/skill/SKILL.md

@@ -83,6 +83,18 @@ gno search "your query"               # BM25 keyword search
 --no-pager        Disable paging
 ```
 
+## Important: Embedding After Changes
+
+If you edit/create files that should be searchable via vector search:
+
+```bash
+gno index              # Full re-index (sync + embed)
+# or
+gno embed              # Embed only (if already synced)
+```
+
+MCP `gno.sync` and `gno.capture` do NOT auto-embed. Use CLI for embedding.
+
 ## Reference Documentation
 
 | Topic                                                 | File                                 |

package/assets/skill/mcp-reference.md

@@ -1,8 +1,23 @@
-# GNO MCP Reference
+# GNO MCP Installation
 
-GNO provides an MCP (Model Context Protocol) server for AI integration.
+GNO provides an MCP (Model Context Protocol) server for AI client integration.
 
-## Setup
+> **Full reference**: See [gno.sh/docs/MCP](https://www.gno.sh/docs/MCP) for complete tool documentation.
+
+## Quick Install
+
+```bash
+# Claude Desktop (default)
+gno mcp install
+
+# Claude Code
+gno mcp install -t claude-code
+
+# With write tools enabled
+gno mcp install --enable-write
+```
+
+## Manual Setup
 
 ### Claude Desktop
 
@@ -19,202 +34,28 @@ Add to `claude_desktop_config.json`:
 }
 ```
 
-Config location:
+Config locations:
 
 - macOS: `~/Library/Application Support/Claude/claude_desktop_config.json`
 - Windows: `%APPDATA%\Claude\claude_desktop_config.json`
 - Linux: `~/.config/Claude/claude_desktop_config.json`
 
-### Start Server
+### Claude Code
 
 ```bash
-gno mcp
-```
-
-Runs JSON-RPC 2.0 over stdio.
-
-## Tools
-
-### gno.search
-
-BM25 keyword search.
-
-```json
-{
-  "query": "search terms",
-  "collection": "optional-collection",
-  "limit": 5,
-  "minScore": 0.5,
-  "lang": "en",
-  "tagsAny": ["project", "work"],
-  "tagsAll": ["reviewed"]
-}
-```
-
-| Parameter    | Description                         |
-| ------------ | ----------------------------------- |
-| `query`      | Search query (required)             |
-| `collection` | Filter by collection                |
-| `limit`      | Max results (default: 5)            |
-| `minScore`   | Minimum score 0-1                   |
-| `tagsAny`    | Filter: has ANY of these tags (OR)  |
-| `tagsAll`    | Filter: has ALL of these tags (AND) |
-
-### gno.vsearch
-
-Vector semantic search. Same parameters as `gno.search`.
-
-### gno.query
-
-Hybrid search (best quality).
-
-```json
-{
-  "query": "search terms",
-  "collection": "optional-collection",
-  "limit": 5
-}
-```
-
-**Search modes** (via parameters):
-
-| Mode     | Parameters       | Time  |
-| -------- | ---------------- | ----- |
-| Fast     | `fast: true`     | ~0.7s |
-| Default  | (none)           | ~2-3s |
-| Thorough | `thorough: true` | ~5-8s |
-
-Default skips expansion, with reranking. Use `thorough: true` for best recall.
-
-**Agent retry strategy**: Use default mode first. If no relevant results:
-
-1. Rephrase the query (free, often effective)
-2. Then try `thorough: true` for better recall
-
-### gno.get
-
-Retrieve document by reference.
-
-```json
-{
-  "ref": "gno://collection/path or #docid",
-  "fromLine": 1,
-  "lineCount": 100,
-  "lineNumbers": true
-}
-```
-
-### gno.multi_get
-
-Retrieve multiple documents.
-
-```json
-{
-  "refs": ["gno://work/doc1.md", "#a1b2c3d4"],
-  "maxBytes": 10240,
-  "lineNumbers": true
-}
-```
-
-Or by pattern:
-
-```json
-{
-  "pattern": "work/**/*.md",
-  "maxBytes": 10240
-}
-```
-
-### gno.status
-
-Get index status.
-
-```json
-{}
-```
-
-### gno.list_tags
-
-List tags with document counts.
-
-```json
-{
-  "collection": "optional-collection",
-  "prefix": "project/"
-}
-```
-
-| Parameter    | Description                             |
-| ------------ | --------------------------------------- |
-| `collection` | Filter by collection                    |
-| `prefix`     | Filter by tag prefix (e.g., `project/`) |
-
-### gno.tag
-
-Add or remove tag from document.
-
-```json
-{
-  "ref": "gno://work/readme.md",
-  "tag": "project/api",
-  "action": "add"
-}
+gno mcp install -t claude-code -s user    # User scope
+gno mcp install -t claude-code -s project # Project scope
 ```
 
-| Parameter | Description                        |
-| --------- | ---------------------------------- |
-| `ref`     | Document URI or docid (required)   |
-| `tag`     | Tag string (required)              |
-| `action`  | `add` or `remove` (default: `add`) |
+## Check Status
 
-## Resources
-
-Documents accessible as MCP resources:
-
-```
-gno://{collection}/{path}
-```
-
-Examples:
-
-- `gno://work/contracts/nda.docx`
-- `gno://notes/2025/01/meeting.md`
-
-Returns Markdown content with line numbers.
-
-## Response Format
-
-All tools return:
-
-```json
-{
-  "content": [
-    { "type": "text", "text": "Human-readable summary" }
-  ],
-  "structuredContent": {
-    "results": [...],
-    "meta": { "query": "...", "mode": "hybrid" }
-  }
-}
+```bash
+gno mcp status
 ```
 
-## Error Handling
+## Uninstall
 
-Errors return:
-
-```json
-{
-  "isError": true,
-  "content": [
-    { "type": "text", "text": "Error: Document not found" }
-  ]
-}
+```bash
+gno mcp uninstall
+gno mcp uninstall -t claude-code
 ```
-
-## Graceful Degradation
-
-`gno.query` degrades gracefully:
-
-- No vectors → BM25 only
-- No expansion model → skips expansion
-- No rerank model → skips reranking

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@gmickel/gno",
-  "version": "0.10.4",
+  "version": "0.12.0",
   "description": "Local semantic search for your documents. Index Markdown, PDF, and Office files with hybrid BM25 + vector search.",
   "keywords": [
     "embeddings",

package/src/cli/commands/embed.ts

@@ -161,13 +161,12 @@ async function processBatches(ctx: BatchContext): Promise<BatchResult> {
       continue;
     }
 
-    // Store vectors
+    // Store vectors (embeddedAt set by DB)
     const vectors: VectorRow[] = batch.map((b, idx) => ({
       mirrorHash: b.mirrorHash,
       seq: b.seq,
       model: ctx.modelUri,
       embedding: new Float32Array(embeddings[idx] as number[]),
-      embeddedAt: new Date().toISOString(),
     }));
 
     const storeResult = await ctx.vectorIndex.upsertVectors(vectors);
@@ -373,6 +372,26 @@ export async function embed(options: EmbedOptions = {}): Promise<EmbedResult> {
       return { success: false, error: result.error };
     }
 
+    // Sync vec index if any vec0 writes failed (matches embedBacklog behavior)
+    if (vectorIndex.vecDirty) {
+      const syncResult = await vectorIndex.syncVecIndex();
+      if (syncResult.ok) {
+        const { added, removed } = syncResult.value;
+        if (added > 0 || removed > 0) {
+          if (!options.json) {
+            process.stdout.write(
+              `\n[vec] Synced index: +${added} -${removed}\n`
+            );
+          }
+        }
+        vectorIndex.vecDirty = false;
+      } else if (!options.json) {
+        process.stdout.write(
+          `\n[vec] Sync failed: ${syncResult.error.message}\n`
+        );
+      }
+    }
+
     return {
       success: true,
       embedded: result.embedded,