hablas-ai 2.0.1 → 2.0.3
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 +16 -0
- package/dist/index.js +184 -123
- package/package.json +1 -1
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,21 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## [2.0.3] — 2026-06-12 · Prompt Strength Rebuild
|
|
4
|
+
|
|
5
|
+
### Changed
|
|
6
|
+
- Rebuilt the single-agent system prompt using the strongest parts of the old specialist prompts.
|
|
7
|
+
- Folded product, architecture, engineering, debugging, research, review, and devops behavior into one unified Hablas prompt.
|
|
8
|
+
- Strengthened tool-discipline rules for implementation work.
|
|
9
|
+
- Added stronger anti-fake-completion behavior in the prompt so implementation tasks are expected to use real workspace actions.
|
|
10
|
+
|
|
11
|
+
## [2.0.2] — 2026-06-12 · Implementation Routing and Tool-Enforcement Fix
|
|
12
|
+
|
|
13
|
+
### Fixed
|
|
14
|
+
- Implementation requests are now classified more robustly, including common typos like `craet`.
|
|
15
|
+
- Architecture-only requests no longer get misrouted into implementation just because they mention engineering context.
|
|
16
|
+
- Implementation turns now explicitly require real workspace tool usage.
|
|
17
|
+
- If the model tries to answer an implementation task with inline code only and no tool calls, Hablas now forces a second attempt through the tool path instead of accepting the fake completion.
|
|
18
|
+
|
|
3
19
|
## [2.0.1] — 2026-06-12 · Setup Flow Restore
|
|
4
20
|
|
|
5
21
|
### Changed
|