agent-working-memory 0.15.4 → 0.15.5

package/README.md

@@ -135,6 +135,57 @@ and whether anything needs a re-run of setup.
 
 ---
 
+## Upgrading
+
+```bash
+npm install -g agent-working-memory@latest
+```
+
+**That alone is not enough** — it updates the engine, not the wiring. Do the second step for
+however you installed it:
+
+| Installed as | Second step |
+|---|---|
+| CLI (`awm setup`) | `awm setup --global` again. It upgrades in place: your database, agent id, ports and any env you set are all preserved |
+| Claude Code plugin | `/plugin marketplace update agent-working-memory` then `/plugin install awm@agent-working-memory` |
+| Claude Desktop | Rebuild the `.mcpb` and re-install it |
+
+Then **restart Claude Code** (or Desktop). A running MCP connection keeps the code it loaded
+at spawn time, so nothing changes until a fresh session starts.
+
+**A plugin marketplace installed from a local directory does not refresh itself.** Measured:
+a GitHub-sourced marketplace on one machine refreshed on a schedule, while a directory-sourced
+one had not moved in 167 days. Since the npm route registers a directory marketplace, the
+`/plugin marketplace update` line above is required, not optional — skip it and you get new
+engine code with old hooks, which half-works and is confusing for it.
+
+```
+memory_whoami
+```
+
+reports the version, store and agent the running process **actually** has. It is the fastest
+way to confirm an upgrade landed, and the only one that cannot be fooled by a stale cache.
+
+Your memories are not touched by any of this. Every release so far has been backward
+compatible, and the store is the same SQLite file before and after.
+
+---
+
+## What's new — v0.15.5
+
+| | |
+|---|---|
+| **0.15.x** | Runs as a **Claude Code plugin** and a **Claude Desktop extension**, both sharing one store with the CLI. Installs from npm with no GitHub access — `awm plugin` prints the two lines to paste. Every write now records which client produced it (`client=claude-code`) |
+| **0.14.6** | `awm setup` caught up with four releases of engine changes it had silently fallen behind. Added the release checks that now block on that kind of drift |
+| **0.14.5** | Benchmark instrument corrected: the eval had been querying as the wrong agent, so a product feature scored as a ranking failure. The real identifier baseline is **92.7%**, not 68% |
+| **0.14.2** | Each session's sidecar binds the first free port instead of giving up, so several projects can run at once |
+| **0.14.1** | An empty recall says `RECALL ABSTAINED` with the withheld count, rather than claiming the memory is absent |
+
+Full history, including every benchmark figure that was corrected and why:
+[CHANGELOG.md](CHANGELOG.md).
+
+---
+
 ## Why it works when a vector store doesn't
 
 Most "memory for AI" stores everything and retrieves by similarity. AWM makes the opposite bet:
@@ -243,7 +294,7 @@ npm run test:linux      # build + full suite on Linux (781/781)
 
 ## Status
 
-Active development, v0.15.4. Core retrieval, consolidation, MCP integration, hooks, task
+Active development, v0.15.5. Core retrieval, consolidation, MCP integration, hooks, task
 management, and the HTTP API are stable and in daily production use. PGlite backend stable;
 networked Postgres experimental.
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agent-working-memory",
-  "version": "0.15.4",
+  "version": "0.15.5",
   "description": "Persistent, local-first memory for AI agents. Selective by design: filters what's worth keeping, links what's related, and says nothing when it doesn't know. SQLite + ONNX, no cloud, no API keys.",
   "keywords": [
     "ai-agents",

package/plugin/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "awm",
   "displayName": "Agent Working Memory",
-  "version": "0.15.4",
+  "version": "0.15.5",
   "description": "Persistent, local-first memory for AI agents. Selective by design: filters what's worth keeping, links what's related, and says nothing when it doesn't know. SQLite + ONNX, no cloud, no API keys.",
   "author": {
     "name": "Robert Winter"

package/plugin/README.md

@@ -22,7 +22,7 @@ The launcher defaults `AWM_DB_PATH` to `~/.awm/memory.db` — **the same file
 than quietly starting an empty one. Set `AWM_DB_PATH` yourself to override.
 
 It prefers an installed copy of the package (global, or a local `node_modules`) and falls
-back to `npx -p agent-working-memory@0.15.4`. The npx path rebuilds
+back to `npx -p agent-working-memory@0.15.5`. The npx path rebuilds
 `better-sqlite3` on a cold cache, which is slow enough to look like a hang, so a real
 install is worth having:
 

package/plugin/bin/awm-mcp-launcher.cjs

@@ -6,7 +6,7 @@
 //
 // This is a REAL source file, not a string inside the generator, because it is real code:
 // it gets syntax-checked, linted and tested like everything else. `npm run build:plugin`
-// copies it to plugin/bin/ and substitutes 0.15.4.
+// copies it to plugin/bin/ and substitutes 0.15.5.
 //
 // Two jobs a bare `mcpServers` entry in plugin.json cannot do:
 //
@@ -95,7 +95,7 @@ if (found) {
   const isWin = process.platform === 'win32';
   const child = spawn(
     isWin ? 'npx.cmd' : 'npx',
-    ['-y', '-p', 'agent-working-memory@0.15.4', 'awm', 'mcp'],
+    ['-y', '-p', 'agent-working-memory@0.15.5', 'awm', 'mcp'],
     { stdio: 'inherit', env: process.env, shell: isWin }
   );
   child.on('error', (e) => {

package/plugin/skills/awm-memory/SKILL.md

@@ -10,7 +10,7 @@ description: >-
 
 # Agent Working Memory
 
-Generated from `src/adapters/common.ts` (AWM v0.15.4) — the same guidance
+Generated from `src/adapters/common.ts` (AWM v0.15.5) — the same guidance
 `awm setup` writes into CLAUDE.md. Do not edit here; edit the source and re-run
 `npm run build:plugin`.