memgineering 0.7.0 → 0.7.1
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 +25 -1
- package/assets/MEMGINEERING.md +1 -1
- package/dist/index.js +4 -1
- package/package.json +1 -1
package/CHANGELOG.md
CHANGED
|
@@ -11,6 +11,30 @@ language the reader wants. The bilingual rule the monorepo applies to
|
|
|
11
11
|
|
|
12
12
|
## [Unreleased]
|
|
13
13
|
|
|
14
|
+
## [0.7.1] — 2026-08-17
|
|
15
|
+
|
|
16
|
+
### Changed
|
|
17
|
+
|
|
18
|
+
- **The semantic-search figures 0.7.0 quoted were measured on the set it was
|
|
19
|
+
tuned against.** An independent set of 70 questions — written by someone
|
|
20
|
+
reading only the notes and never a search result, with the expected note fixed
|
|
21
|
+
from what they had just read — scores lower on the same brain and the same
|
|
22
|
+
server: the right note is in the top five for **72.9%** of them and ranked
|
|
23
|
+
first for **45.7%**, and for a question asked across a language, **61.1%** and
|
|
24
|
+
**33.3%**. Text search alone scores 47.1% / 35.7% overall and 5.6% / 0% across
|
|
25
|
+
a language on that set, so meaning still moves every category it touches, and
|
|
26
|
+
cross-language remains the largest move either set shows. Both sets are real
|
|
27
|
+
measurements; only the tuned one had been reported. Expect the lower pair from
|
|
28
|
+
a question nobody tuned for.
|
|
29
|
+
|
|
30
|
+
### Fixed
|
|
31
|
+
|
|
32
|
+
- `recall --help` no longer offers `--detail summary`, `chunks`, and `full` as
|
|
33
|
+
though a hosted brain could answer them. It now says which two levels a hosted
|
|
34
|
+
brain does answer, and points at a brain in a folder or `open <handle>` for the
|
|
35
|
+
deeper ones. The refusal was already correct; the help text described a
|
|
36
|
+
different product.
|
|
37
|
+
|
|
14
38
|
## [0.7.0] — 2026-08-16
|
|
15
39
|
|
|
16
40
|
Recall reads what a note says, not only what it is called. And the pile of
|
|
@@ -181,6 +205,7 @@ that goes one approved change at a time.
|
|
|
181
205
|
the shape of its answers without changing its name.
|
|
182
206
|
|
|
183
207
|
### Fixed
|
|
208
|
+
|
|
184
209
|
- **`push` now reports the notes it never sent.** A note held back by the name
|
|
185
210
|
deny list or by an ignore rule was dropped before the upload and appeared in
|
|
186
211
|
no summary — a 277-note brain reported `276 note(s) uploaded` and said nothing
|
|
@@ -205,7 +230,6 @@ that goes one approved change at a time.
|
|
|
205
230
|
held each note, instead of pointing at `memgineering reindex` — which reports
|
|
206
231
|
the name and content screens and says nothing about an ignore rule.
|
|
207
232
|
|
|
208
|
-
|
|
209
233
|
- **A `supersedes:` written as a single value is no longer ignored.**
|
|
210
234
|
`supersedes: mem-abc` is what somebody writes when one memory replaces one
|
|
211
235
|
other, and it parsed to nothing at all — no relation, no complaint, no way to
|
package/assets/MEMGINEERING.md
CHANGED
|
@@ -2,7 +2,7 @@
|
|
|
2
2
|
name: memgineering
|
|
3
3
|
description: Use whenever the user refers to something they told you before, asks what was decided, tells you something worth keeping, or settles something that should hold next time. The memory lives in their own folder and outlives this session; check it before answering from guesswork, and write to it when you learn something durable.
|
|
4
4
|
type: skill
|
|
5
|
-
version: 0.7.
|
|
5
|
+
version: 0.7.1
|
|
6
6
|
---
|
|
7
7
|
|
|
8
8
|
# memgineering
|
package/dist/index.js
CHANGED
|
@@ -12424,7 +12424,10 @@ init_errors();
|
|
|
12424
12424
|
init_vault();
|
|
12425
12425
|
init_ui();
|
|
12426
12426
|
function recallCommand() {
|
|
12427
|
-
return new Command18("recall").description("recall memory cards for a question").argument("<query>", "what you want to remember").option("--vault <path>", "which brain to search").option("--local", "search the brain on this machine, even when signed in to a hosted one").option("--scope <scope>", "restrict to one scope").option("--limit <n>", "max cards", (v) => parseInt(v, 10), 8).option(
|
|
12427
|
+
return new Command18("recall").description("recall memory cards for a question").argument("<query>", "what you want to remember").option("--vault <path>", "which brain to search").option("--local", "search the brain on this machine, even when signed in to a hosted one").option("--scope <scope>", "restrict to one scope").option("--limit <n>", "max cards", (v) => parseInt(v, 10), 8).option(
|
|
12428
|
+
"--detail <level>",
|
|
12429
|
+
"title | card | summary | chunks | full (default: card) \u2014 a hosted brain answers title and card only; deeper needs a brain in a folder, or `open <handle>`"
|
|
12430
|
+
).option("--depth <depth>", "exact | standard (default: standard, which also searches bodies)").option("--budget <tokens>", "max estimated size of the answer", (v) => parseInt(v, 10)).action(
|
|
12428
12431
|
async (query, opts) => {
|
|
12429
12432
|
const detail = parseDetail(opts.detail, "card");
|
|
12430
12433
|
const depth = parseDepth(opts.depth);
|
package/package.json
CHANGED