hablas-ai 2.2.5 → 2.2.7
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 +37 -11
- package/dist/index.js +164 -387
- package/package.json +4 -3
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,42 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## [2.2.7] — 2026-06-14 · Stronger Core Agent Semantics and Continuity
|
|
4
|
+
|
|
5
|
+
### Added
|
|
6
|
+
- model-assisted semantic task profiling as the primary turn-classification path
|
|
7
|
+
- session continuity regression test to ensure system-prompt refresh does not wipe conversation history
|
|
8
|
+
|
|
9
|
+
### Changed
|
|
10
|
+
- the core system prompt now has a stronger execution architecture, clearer decision priorities, and a warmer but still technical interaction contract
|
|
11
|
+
- task preparation now uses recent conversation history during profiling instead of relying only on keyword-style routing
|
|
12
|
+
- `updateSystemPrompt()` now preserves normal user, assistant, and tool history instead of collapsing the turn context
|
|
13
|
+
|
|
14
|
+
### Notes
|
|
15
|
+
- Hablas remains a single visible agent
|
|
16
|
+
- vision remains inside image tooling only
|
|
17
|
+
- tests and build pass for this release candidate
|
|
18
|
+
- `npm run lint` still reports pre-existing TypeScript issues outside this patch scope
|
|
19
|
+
|
|
20
|
+
## [2.2.6] — 2026-06-13 · Vision Inside Image Tool (Phase 1)
|
|
21
|
+
|
|
22
|
+
### Added
|
|
23
|
+
- internal vision support inside `inspect_image`
|
|
24
|
+
- primary and fallback vision model configuration in the runtime config
|
|
25
|
+
- merged vision verdict flow for image relevance and recommended use
|
|
26
|
+
- `/about` now includes Abdulmoin Hablas and the portfolio link again
|
|
27
|
+
|
|
28
|
+
### Changed
|
|
29
|
+
- image tooling now follows a stronger architecture:
|
|
30
|
+
- `search_image_candidates`
|
|
31
|
+
- `inspect_image`
|
|
32
|
+
- `download_asset`
|
|
33
|
+
- `download_asset` can recover direct image candidates from HTML pages before failing
|
|
34
|
+
- image-related prompt guidance is stronger and more architectural
|
|
35
|
+
|
|
36
|
+
### Notes
|
|
37
|
+
This is the first integrated vision phase.
|
|
38
|
+
Vision is a capability of the image tool layer only, not a separate assistant runtime.
|
|
39
|
+
|
|
3
40
|
## [2.2.5] — 2026-06-13 · Models Pagination and Startup Redesign
|
|
4
41
|
|
|
5
42
|
### Added
|
|
@@ -22,14 +59,3 @@
|
|
|
22
59
|
This release completes backlog items:
|
|
23
60
|
- M-001 — models pagination / browsing UX
|
|
24
61
|
- M-002 — startup screen / first-view redesign
|
|
25
|
-
|
|
26
|
-
## [2.2.4] — 2026-06-12 · Response Presentation and Personality Refinement
|
|
27
|
-
|
|
28
|
-
### Changed
|
|
29
|
-
- Improved final response presentation toward a more executive structure.
|
|
30
|
-
- Assistant markdown rendering is now richer, including better table rendering and stronger section visibility.
|
|
31
|
-
- Hablas personality was warmed up without turning fluffy: still technical and direct, but less cold and more partner-like.
|
|
32
|
-
- Final-response contract in the system prompt now explicitly prefers clearer completion sections and better next-step suggestions.
|
|
33
|
-
|
|
34
|
-
### Added
|
|
35
|
-
- formal improvement backlog document for tracked UX and personality issues.
|