agenr 0.9.3 → 0.9.5
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 +26 -0
- package/dist/chunk-AJR3OWNP.js +1912 -0
- package/dist/chunk-IC7FMYA3.js +2147 -0
- package/dist/chunk-NUFBXKLH.js +1912 -0
- package/dist/chunk-RMYAYY4E.js +1908 -0
- package/dist/chunk-RYFTWDLF.js +2147 -0
- package/dist/cli-main.d.ts +5 -6
- package/dist/cli-main.js +410 -501
- package/dist/openclaw-plugin/index.js +392 -24
- package/package.json +1 -1
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,31 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## 0.9.5 - 2026-02-27
|
|
4
|
+
|
|
5
|
+
### Added
|
|
6
|
+
- Feedback-driven recall scoring (#267 Phase 1): recalled entries are tracked per
|
|
7
|
+
session and compared against agent responses at session end. Entries that are
|
|
8
|
+
used get quality score boosts; unused entries drift slightly downward.
|
|
9
|
+
- Correction signal: if the agent stores a contradicting entry via agenr_store
|
|
10
|
+
during the session, the recalled entry that was corrected gets a strong
|
|
11
|
+
negative signal (quality score drops toward 0).
|
|
12
|
+
- Rolling quality_score (0-1) per entry, integrated into recall ranking formula.
|
|
13
|
+
Consistently useful entries rank higher over time.
|
|
14
|
+
- Entry-type quality floor: facts and preferences cannot drop below 0.35 to
|
|
15
|
+
prevent background-context entries from being unfairly penalized.
|
|
16
|
+
- Auto-strengthen: entries reaching recall count milestones (3, 10, 25) get
|
|
17
|
+
importance bumped by 1 (capped at 9, never auto-promotes to 10).
|
|
18
|
+
- Quality score distribution in `agenr health` output.
|
|
19
|
+
|
|
20
|
+
## 0.9.4 (2026-02-27)
|
|
21
|
+
|
|
22
|
+
### Changed
|
|
23
|
+
- refactor: removed top-level `model` config field; `models` is now required with all four task keys (extraction, claimExtraction, contradictionJudge, handoffSummary) always explicit (#277)
|
|
24
|
+
- `resolveModelForTask` simplified to direct lookup (no fallback chain)
|
|
25
|
+
- `isCompleteConfig` now checks for complete `models` instead of top-level `model`
|
|
26
|
+
- Old configs with top-level `model` auto-upgrade on read (value populates all task models)
|
|
27
|
+
- `config set model <value>` removed; use `config set models.extraction <value>` etc.
|
|
28
|
+
|
|
3
29
|
## 0.9.3 (2026-02-26)
|
|
4
30
|
|
|
5
31
|
### Added
|