minnimemory 1.3.1 → 1.3.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +36 -0
- package/README.md +2 -1
- package/package.json +1 -1
- package/pkg/chunks/chunk-LVFL6LTT.js +156 -0
- package/pkg/chunks/chunk-UXV5KSVU.js +52 -0
- package/pkg/chunks/{mcpServer-IWTKZFHK.js → mcpServer-3WQFI5UL.js} +1 -1
- package/pkg/cli.js +3 -3
- package/pkg/index.js +1 -1
- package/pkg/chunks/chunk-2HVUP6MF.js +0 -156
- package/pkg/chunks/chunk-2ZVFFYPQ.js +0 -50
package/CHANGELOG.md
CHANGED
|
@@ -3,6 +3,42 @@
|
|
|
3
3
|
All notable changes to the `minnimemory` package. Dates are publish dates; an entry that has not
|
|
4
4
|
been published yet says `unreleased` and gets its date on the day it goes out.
|
|
5
5
|
|
|
6
|
+
## 1.3.3 - 2026-10-13
|
|
7
|
+
|
|
8
|
+
- Fix: the recall hook `install` registers printed nothing once `apply` had compiled the memory
|
|
9
|
+
folder. With no `--dir`, the hook takes the folder `init` recorded and always ranked it as an
|
|
10
|
+
uncompiled folder, which finds no match in a compiled one; it now checks the phase the way
|
|
11
|
+
`--dir` does. Effect in a live Claude Code session with the install kept (a 30-task memory
|
|
12
|
+
benchmark, 2 repeats): +20.3% per session against the unoptimized memory before the fix,
|
|
13
|
+
-27.3% after.
|
|
14
|
+
|
|
15
|
+
## 1.3.2 - 2026-10-13
|
|
16
|
+
|
|
17
|
+
- Fix (#13): a sorted `AlwaysOnMemory/` file that no file Claude Code loads imports any more
|
|
18
|
+
(the import line was removed, or the user `CLAUDE.md` is gone) stayed labelled always-on:
|
|
19
|
+
`doctor` said "nothing to apply" and the index header still told the agent the file was
|
|
20
|
+
loaded, so it was never opened. `doctor` now reports it (MM010, with the import line to put
|
|
21
|
+
back or the move to `OnDemandMemory/` to make), the drift notice says so once with that
|
|
22
|
+
remedy, and the summary gets a "not loaded:" line. Nothing moves on its own: the import may
|
|
23
|
+
have gone on purpose (DESIGN.md 6.10, decision 25). Raised only when the loaders are known (a
|
|
24
|
+
saved set, an import that still reaches the folder, or a host an earlier apply rewrote is in
|
|
25
|
+
the report), so an incomplete agent report does not raise it.
|
|
26
|
+
- Changed (#14): over MCP, `init`/`doctor` on a workspace with nothing to apply lead with that
|
|
27
|
+
verdict and give the audit as one line per rule (count and worst severity) instead of up to
|
|
28
|
+
four findings per rule; `format: "json"` still carries every finding. On an 85-file memory
|
|
29
|
+
folder the weekly `doctor` reply went from about 1,400 to about 350 tokens (approx-v2). The terminal
|
|
30
|
+
output is unchanged.
|
|
31
|
+
- Changed (#14): `apply(content: true)` no longer sends a backup's content, which is the source
|
|
32
|
+
file as it is on disk: backups come as `copies` (from -> to) for the agent to perform before
|
|
33
|
+
the writes, and `files` holds the writes only. The JSON form of `apply`/`undo` is compact, as
|
|
34
|
+
`doctor`'s already was. On the same 85-file folder the reply went from about 16,300 to
|
|
35
|
+
about 14,500 tokens (text) and 26,900 to 24,400 (json).
|
|
36
|
+
- Changed (#14): on a sorted memory folder, `doctor`'s summary now ends with the same "sorted
|
|
37
|
+
memory folder: N routed files" and "drift: N files changed" lines a project root gets, so
|
|
38
|
+
day-two drift is not only MM010 lines among the audit findings; and a re-apply whose verdict
|
|
39
|
+
is "sort" no longer says "skip it if tokens are all you want", since it also brings the index
|
|
40
|
+
and manifest back in step.
|
|
41
|
+
|
|
6
42
|
## 1.3.1 - 2026-10-12
|
|
7
43
|
|
|
8
44
|
- Fixed: on a memory folder with topic files the index never linked, `doctor` after `apply`
|
package/README.md
CHANGED
|
@@ -225,7 +225,8 @@ server tells the agent the loop on connect, so these four requests are all it ta
|
|
|
225
225
|
Every write goes through your own tool's approval prompt. The MCP server itself only reads, and
|
|
226
226
|
it never sends your files' content through the model: `apply`/`undo` return the command and the
|
|
227
227
|
operation list (about 240 tokens on the fixture above, about 300 on an 85-file memory folder);
|
|
228
|
-
`content: true` adds the files
|
|
228
|
+
`content: true` adds the files to write, for an agent that cannot run commands; backups are
|
|
229
|
+
listed as copies to make (from -> to), since their content is the file already on disk.
|
|
229
230
|
`init`/`doctor` return a summary diff over MCP by default (the plan table plus one line per
|
|
230
231
|
file); pass `diff: "full"` for the complete unified diff the terminal always shows.
|
|
231
232
|
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "minnimemory",
|
|
3
|
-
"version": "1.3.
|
|
3
|
+
"version": "1.3.3",
|
|
4
4
|
"description": "MinniMemoryMCP: an MCP server that gives an AI coding agent curated memory. Tools audit the agent's memory file, compile it into a small AlwaysOnMemory body plus OnDemandMemory files, serve it back on demand and keep it maintained. Deterministic, lossless, offline. Built by MinniAI.",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"mcpName": "io.github.MinniAI-com/minnimemory",
|