probe-research 0.10.3 → 0.10.5
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/bin/probe-research.js +14 -3
- package/package.json +1 -1
package/bin/probe-research.js
CHANGED
|
@@ -35,10 +35,21 @@ const DIST = "probe-research";
|
|
|
35
35
|
* `probe-research==<our version>` resolves to a PyPI version that does not
|
|
36
36
|
* exist. That is exactly how 0.10.1 shipped broken.
|
|
37
37
|
*
|
|
38
|
-
*
|
|
39
|
-
*
|
|
38
|
+
* Three reasons to bump it, and only three:
|
|
39
|
+
* 1. the launcher starts depending on a NEW CLI feature;
|
|
40
|
+
* 2. the plugin's skill prose starts referencing new CLI verbs, so a fresh
|
|
41
|
+
* install can never pair old-CLI with new-prose (0066 tags review);
|
|
42
|
+
* 3. a shipped CLI version is KNOWN BROKEN and this is the repair path.
|
|
43
|
+
*
|
|
44
|
+
* (3) is why the floor is 0.27.1. 0.26.0 through 0.27.0 crash on every fresh
|
|
45
|
+
* `probe wizard` — a KeyError the moment the plan names auto-update (#120) —
|
|
46
|
+
* and they all cleared the previous 0.24.0 floor, so this launcher happily
|
|
47
|
+
* handed users straight to a CLI that could not complete a setup. `npx
|
|
48
|
+
* probe-research` IS the advertised from-zero entry point; it has to be able
|
|
49
|
+
* to repair the thing it installed. A healthy 0.26.x install pays one refetch,
|
|
50
|
+
* which is the right trade against a broken one staying broken forever.
|
|
40
51
|
*/
|
|
41
|
-
const MIN_CLI = "0.
|
|
52
|
+
const MIN_CLI = "0.27.1";
|
|
42
53
|
const UV_INSTALL = "curl -LsSf https://astral.sh/uv/install.sh | sh";
|
|
43
54
|
|
|
44
55
|
/** Compare dotted numeric versions. Returns true when `a` >= `b`. */
|