@uluops/registry-mcp 0.11.2 → 0.11.4
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
CHANGED
|
@@ -1,5 +1,25 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## [0.11.4] - 2026-10-03
|
|
4
|
+
|
|
5
|
+
### Changed
|
|
6
|
+
|
|
7
|
+
- **Both caveats no longer say the pooled figures pool every version.** 0.11.3 said health, failure-domain and taxonomy figures "pool every version of the definition", and the pooled-versions caveat on `get_effectiveness` and `get_health` said "these figures pool every version". The issue-derived figures do not: registry-api `getEffectiveness` filters issues to the `definitionId` of whichever run carries one first (tracker `7ed97aa5`), so the false-positive, declined and resolution rates, the taxonomy, and the health factors built from them come from one version that need not be the one requested; and its runs carry no org filter (definition-version-dispositions A35 review, tracker run #67, amendment AS). The fix for `7ed97aa5` is not chosen, so the new wording uses "may" and "possibly" and stays true before and after it. Execution counts are now named among the figures that are not per-version (amendment AN). **What changes for you:** description and response-block text only.
|
|
8
|
+
|
|
9
|
+
### Fixed
|
|
10
|
+
|
|
11
|
+
- **The pooled-versions caveat is now pinned word for word**, like the cross-version one (it was asserted only by `toContain` against descriptions and responses, which a reworded constant passes).
|
|
12
|
+
|
|
13
|
+
## [0.11.3] - 2026-10-02
|
|
14
|
+
|
|
15
|
+
### Changed
|
|
16
|
+
|
|
17
|
+
- **Cross-version caveat reworded again: it no longer says every figure is per-version.** 0.11.2 said "each version ran in its own period, on its own artifacts" of every figure in `get_lineage`, `get_evolution`, `compare_effectiveness`, `get_diff_impact` and `get_translation_analytics`. Health, failure-domain and taxonomy figures are not per-version: the registry computes them over every version and persists them under whichever version was requested, so lineage, evolution and compare lay one pooled figure, stamped per version, side by side, and a difference there reflects when each version was last recomputed (definition-version-dispositions A33 review, tracker run #66; verified in registry-api `getEffectiveness` and `definition-analytics-service`). The new sentence says so, keeps the per-version clause for pass rates and scores, and drops "grades", which the dispositions spec reserves for another sense. **What changes for you:** description and response-block text only.
|
|
18
|
+
|
|
19
|
+
### Fixed
|
|
20
|
+
|
|
21
|
+
- **The test now pins the full sentence**, and `@uluops/ops-mcp`'s test pins the same literal, so a one-word drift in either package fails (A33 test-architect: "artifacts" → "artifact" survived 0.11.2's partial pin).
|
|
22
|
+
|
|
3
23
|
## [0.11.2] - 2026-10-02
|
|
4
24
|
|
|
5
25
|
### Changed
|
|
@@ -19,6 +19,20 @@
|
|
|
19
19
|
* Why the wording names "alone or combined": 0.11.1 said "from these figures alone",
|
|
20
20
|
* which licensed a ranking as soon as any second source (such as a dispositions
|
|
21
21
|
* record) was added (A31 review, run #65).
|
|
22
|
+
*
|
|
23
|
+
* Why it names the pooled figures: 0.11.2 said "each version ran in its own period, on its own artifacts" of
|
|
24
|
+
* every figure. Health, failure-domain and taxonomy figures are computed over every version and persisted under
|
|
25
|
+
* whichever version was requested (registry-api getEffectiveness → definition_metrics), so lineage, evolution and
|
|
26
|
+
* compare show one pooled figure stamped per version; a difference there reflects recompute time, not the version
|
|
27
|
+
* (A33 review, run #66). The sentence is identical in @uluops/registry-mcp and @uluops/ops-mcp; each package's test
|
|
28
|
+
* pins it word for word.
|
|
29
|
+
*
|
|
30
|
+
* Why "may" and "possibly" (0.11.4 / 0.27.1): 0.11.3 / 0.26.2 said these figures "pool every version". The
|
|
31
|
+
* issue-derived parts are not pooled: getEffectiveness filters issues to the definitionId of whichever run carries one
|
|
32
|
+
* first (registry-api analytics-service.ts, tracker issue 7ed97aa5), and its runs carry no org filter. The fix for
|
|
33
|
+
* 7ed97aa5 is not yet chosen, so the sentence is written to stay true both before and after it (A35 review, run #67,
|
|
34
|
+
* amendment AS). Execution counts are named because AN put volumes into the class rule (they are the same pooled
|
|
35
|
+
* count stamped per version row).
|
|
22
36
|
*/
|
|
23
37
|
export declare const CROSS_VERSION_CAVEAT: string;
|
|
24
38
|
export declare const POOLED_VERSIONS_CAVEAT: string;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"cross-version-caveat.d.ts","sourceRoot":"","sources":["../../src/tools/cross-version-caveat.ts"],"names":[],"mappings":"AAAA
|
|
1
|
+
{"version":3,"file":"cross-version-caveat.d.ts","sourceRoot":"","sources":["../../src/tools/cross-version-caveat.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAmCG;AACH,eAAO,MAAM,oBAAoB,QAUiB,CAAC;AAEnD,eAAO,MAAM,sBAAsB,QAK6C,CAAC;AAEjF,8EAA8E;AAC9E,eAAO,MAAM,mBAAmB,oHAMtB,CAAC;AAEX,gFAAgF;AAChF,eAAO,MAAM,qBAAqB,8CAA+C,CAAC"}
|
|
@@ -19,15 +19,36 @@
|
|
|
19
19
|
* Why the wording names "alone or combined": 0.11.1 said "from these figures alone",
|
|
20
20
|
* which licensed a ranking as soon as any second source (such as a dispositions
|
|
21
21
|
* record) was added (A31 review, run #65).
|
|
22
|
+
*
|
|
23
|
+
* Why it names the pooled figures: 0.11.2 said "each version ran in its own period, on its own artifacts" of
|
|
24
|
+
* every figure. Health, failure-domain and taxonomy figures are computed over every version and persisted under
|
|
25
|
+
* whichever version was requested (registry-api getEffectiveness → definition_metrics), so lineage, evolution and
|
|
26
|
+
* compare show one pooled figure stamped per version; a difference there reflects recompute time, not the version
|
|
27
|
+
* (A33 review, run #66). The sentence is identical in @uluops/registry-mcp and @uluops/ops-mcp; each package's test
|
|
28
|
+
* pins it word for word.
|
|
29
|
+
*
|
|
30
|
+
* Why "may" and "possibly" (0.11.4 / 0.27.1): 0.11.3 / 0.26.2 said these figures "pool every version". The
|
|
31
|
+
* issue-derived parts are not pooled: getEffectiveness filters issues to the definitionId of whichever run carries one
|
|
32
|
+
* first (registry-api analytics-service.ts, tracker issue 7ed97aa5), and its runs carry no org filter. The fix for
|
|
33
|
+
* 7ed97aa5 is not yet chosen, so the sentence is written to stay true both before and after it (A35 review, run #67,
|
|
34
|
+
* amendment AS). Execution counts are named because AN put volumes into the class rule (they are the same pooled
|
|
35
|
+
* count stamped per version row).
|
|
22
36
|
*/
|
|
23
|
-
export const CROSS_VERSION_CAVEAT = 'Cross-version figures here are observational
|
|
24
|
-
'and
|
|
25
|
-
'
|
|
26
|
-
'
|
|
27
|
-
'
|
|
28
|
-
|
|
29
|
-
'
|
|
30
|
-
'
|
|
37
|
+
export const CROSS_VERSION_CAVEAT = 'Cross-version figures here are observational. Some may not be per-version at all: health, failure-domain, ' +
|
|
38
|
+
'taxonomy and execution-count figures, where present, may be computed for the definition rather than the ' +
|
|
39
|
+
'version they are shown under (pooling its runs across versions, and possibly across orgs, or taking ' +
|
|
40
|
+
'issue-derived parts from a single version that need not be the one shown) and are stored against a version ' +
|
|
41
|
+
'whenever it is recomputed, so a difference between versions there can reflect when and how each was recomputed ' +
|
|
42
|
+
'rather than the versions. Pass rates and scores, where per-version, come from each version\'s own runs in its ' +
|
|
43
|
+
'own period, on its own artifacts, and are the agent\'s own assessments of those artifacts, so an edit that ' +
|
|
44
|
+
'changes how lenient it is moves them without changing quality. No difference between versions here is evidence ' +
|
|
45
|
+
'that an edit made the definition better or worse; do not rank versions or recommend one on the basis of these ' +
|
|
46
|
+
'figures, alone or combined with other figures.';
|
|
47
|
+
export const POOLED_VERSIONS_CAVEAT = 'These figures are not specific to the requested version: run-based figures pool every version of the ' +
|
|
48
|
+
'definition run in the window (possibly across orgs), and issue-derived figures (false-positive, declined and ' +
|
|
49
|
+
'resolution rates, taxonomy, and the parts of health built from them) may come from a single version that need ' +
|
|
50
|
+
'not be the one requested. They cannot show whether one version differs from another. Do not compare these ' +
|
|
51
|
+
'figures across versions, and do not report them as the requested version\'s.';
|
|
31
52
|
/** Tools whose descriptions and responses must carry CROSS_VERSION_CAVEAT. */
|
|
32
53
|
export const CROSS_VERSION_TOOLS = [
|
|
33
54
|
'get_lineage',
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"cross-version-caveat.js","sourceRoot":"","sources":["../../src/tools/cross-version-caveat.ts"],"names":[],"mappings":"AAAA
|
|
1
|
+
{"version":3,"file":"cross-version-caveat.js","sourceRoot":"","sources":["../../src/tools/cross-version-caveat.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAmCG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAC/B,4GAA4G;IAC5G,0GAA0G;IAC1G,sGAAsG;IACtG,6GAA6G;IAC7G,iHAAiH;IACjH,gHAAgH;IAChH,6GAA6G;IAC7G,iHAAiH;IACjH,gHAAgH;IAChH,gDAAgD,CAAC;AAEnD,MAAM,CAAC,MAAM,sBAAsB,GACjC,uGAAuG;IACvG,+GAA+G;IAC/G,gHAAgH;IAChH,4GAA4G;IAC5G,8EAA8E,CAAC;AAEjF,8EAA8E;AAC9E,MAAM,CAAC,MAAM,mBAAmB,GAAG;IACjC,aAAa;IACb,eAAe;IACf,uBAAuB;IACvB,iBAAiB;IACjB,2BAA2B;CACnB,CAAC;AAEX,gFAAgF;AAChF,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC,mBAAmB,EAAE,YAAY,CAAU,CAAC"}
|