vskill 0.5.47 → 0.5.49
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/dist/commands/add.js +29 -10
- package/dist/commands/add.js.map +1 -1
- package/dist/eval/verdict.d.ts +1 -1
- package/dist/eval/verdict.js +53 -9
- package/dist/eval/verdict.js.map +1 -1
- package/dist/eval-ui/assets/{index-B5TJCTjp.js → index-BjKaIypp.js} +24 -24
- package/dist/eval-ui/index.html +1 -1
- package/dist/installer/migrate.js +11 -7
- package/dist/installer/migrate.js.map +1 -1
- package/dist/security/platform-security.js +18 -4
- package/dist/security/platform-security.js.map +1 -1
- package/package.json +1 -1
package/dist/eval/verdict.js
CHANGED
|
@@ -23,7 +23,8 @@ export function computeVerdict(assertionPassRate, skillRubricAvg, baselineRubric
|
|
|
23
23
|
}
|
|
24
24
|
return "DEGRADING";
|
|
25
25
|
}
|
|
26
|
-
export function verdictExplanation(verdict,
|
|
26
|
+
export function verdictExplanation(verdict, rawScore, rubric) {
|
|
27
|
+
const score = isNaN(rawScore) ? 0 : rawScore;
|
|
27
28
|
const passed = rubric?.filter((r) => r.score >= 0.7) ?? [];
|
|
28
29
|
const failed = rubric?.filter((r) => r.score < 0.4) ?? [];
|
|
29
30
|
const weak = rubric?.filter((r) => r.score >= 0.4 && r.score < 0.7) ?? [];
|
|
@@ -35,7 +36,7 @@ export function verdictExplanation(verdict, score, rubric) {
|
|
|
35
36
|
explanation: `${verdict} (score ${score.toFixed(2)}): evaluation met expectations.${metList}`,
|
|
36
37
|
};
|
|
37
38
|
}
|
|
38
|
-
if (
|
|
39
|
+
if (verdict === "FAIL" || verdict === "DEGRADING") {
|
|
39
40
|
const failedList = failed.length > 0
|
|
40
41
|
? ` Failed criteria: ${failed.map((r) => r.criterion).join(", ")}.`
|
|
41
42
|
: "";
|
|
@@ -51,18 +52,61 @@ export function verdictExplanation(verdict, score, rubric) {
|
|
|
51
52
|
recommendations,
|
|
52
53
|
};
|
|
53
54
|
}
|
|
54
|
-
if (verdict === "INEFFECTIVE"
|
|
55
|
-
|
|
55
|
+
if (verdict === "INEFFECTIVE") {
|
|
56
|
+
if (score < 0.2) {
|
|
57
|
+
const suggestions = [
|
|
58
|
+
...failed.map((r) => `Rework "${r.criterion}" — currently at ${r.score.toFixed(2)}`),
|
|
59
|
+
"Consider adding examples to your system prompt",
|
|
60
|
+
"Review the rubric criteria for achievability",
|
|
61
|
+
];
|
|
62
|
+
return {
|
|
63
|
+
explanation: `${verdict} (score ${score.toFixed(2)}): evaluation is significantly below expectations.`,
|
|
64
|
+
recommendations: suggestions,
|
|
65
|
+
};
|
|
66
|
+
}
|
|
67
|
+
const recommendations = [
|
|
56
68
|
...failed.map((r) => `Rework "${r.criterion}" — currently at ${r.score.toFixed(2)}`),
|
|
57
|
-
|
|
58
|
-
|
|
69
|
+
...weak.map((r) => `Strengthen "${r.criterion}" (score: ${r.score.toFixed(2)})`),
|
|
70
|
+
];
|
|
71
|
+
if (recommendations.length === 0) {
|
|
72
|
+
recommendations.push("Consider restructuring your prompt approach");
|
|
73
|
+
}
|
|
74
|
+
return {
|
|
75
|
+
explanation: `${verdict} (score ${score.toFixed(2)}): below expectations but showing some capability.`,
|
|
76
|
+
recommendations,
|
|
77
|
+
};
|
|
78
|
+
}
|
|
79
|
+
if (verdict === "MARGINAL") {
|
|
80
|
+
const recommendations = [
|
|
81
|
+
...weak.map((r) => `Strengthen "${r.criterion}" (score: ${r.score.toFixed(2)})`),
|
|
82
|
+
...failed.map((r) => `Improve "${r.criterion}" (score: ${r.score.toFixed(2)})`),
|
|
59
83
|
];
|
|
84
|
+
if (recommendations.length === 0) {
|
|
85
|
+
recommendations.push(passed.length > 0
|
|
86
|
+
? "Assertion pass rate is below target despite strong rubric scores — review test case alignment"
|
|
87
|
+
: "Review prompt instructions for areas of improvement");
|
|
88
|
+
}
|
|
60
89
|
return {
|
|
61
|
-
explanation: `${verdict} (score ${score.toFixed(2)}):
|
|
62
|
-
recommendations
|
|
90
|
+
explanation: `${verdict} (score ${score.toFixed(2)}): moderate improvement detected.`,
|
|
91
|
+
recommendations,
|
|
92
|
+
};
|
|
93
|
+
}
|
|
94
|
+
if (verdict === "EMERGING") {
|
|
95
|
+
const recommendations = [
|
|
96
|
+
...failed.map((r) => `Improve "${r.criterion}" (score: ${r.score.toFixed(2)})`),
|
|
97
|
+
...weak.map((r) => `Strengthen "${r.criterion}" (score: ${r.score.toFixed(2)})`),
|
|
98
|
+
];
|
|
99
|
+
if (recommendations.length === 0) {
|
|
100
|
+
recommendations.push(passed.length > 0
|
|
101
|
+
? "Assertion pass rate is below target despite strong rubric scores — review test case alignment"
|
|
102
|
+
: "Add more specific guidance to your prompt instructions");
|
|
103
|
+
}
|
|
104
|
+
return {
|
|
105
|
+
explanation: `${verdict} (score ${score.toFixed(2)}): early promise — focus on weak areas to improve.`,
|
|
106
|
+
recommendations,
|
|
63
107
|
};
|
|
64
108
|
}
|
|
65
|
-
// Default/boundary case (e.g.,
|
|
109
|
+
// Default/boundary case (e.g., PASS/EFFECTIVE with low score)
|
|
66
110
|
const metNote = passed.length > 0
|
|
67
111
|
? ` Passing: ${passed.map((r) => r.criterion).join(", ")}.`
|
|
68
112
|
: "";
|
package/dist/eval/verdict.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"verdict.js","sourceRoot":"","sources":["../../src/eval/verdict.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,mDAAmD;AACnD,8EAA8E;AAI9E,MAAM,UAAU,cAAc,CAC5B,iBAAyB,EACzB,cAAsB,EACtB,iBAAyB,EACzB,mBAA2B,CAAC;IAE5B,MAAM,aAAa,GAAG,iBAAiB,GAAG,gBAAgB,CAAC;IAE3D,+BAA+B;IAC/B,IAAI,iBAAiB,IAAI,GAAG,EAAE,CAAC;QAC7B,IAAI,aAAa,GAAG,IAAI,IAAI,cAAc,GAAG,iBAAiB,GAAG,CAAC,EAAE,CAAC;YACnE,OAAO,WAAW,CAAC;QACrB,CAAC;QACD,OAAO,UAAU,CAAC;IACpB,CAAC;IAED,iCAAiC;IACjC,IAAI,iBAAiB,IAAI,GAAG,EAAE,CAAC;QAC7B,IAAI,aAAa,GAAG,CAAC,EAAE,CAAC;YACtB,OAAO,UAAU,CAAC;QACpB,CAAC;QACD,OAAO,aAAa,CAAC;IACvB,CAAC;IAED,6BAA6B;IAC7B,IAAI,cAAc,GAAG,iBAAiB,EAAE,CAAC;QACvC,OAAO,UAAU,CAAC;IACpB,CAAC;IACD,OAAO,WAAW,CAAC;AACrB,CAAC;AAYD,MAAM,UAAU,kBAAkB,CAChC,OAAsC,EACtC,
|
|
1
|
+
{"version":3,"file":"verdict.js","sourceRoot":"","sources":["../../src/eval/verdict.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,mDAAmD;AACnD,8EAA8E;AAI9E,MAAM,UAAU,cAAc,CAC5B,iBAAyB,EACzB,cAAsB,EACtB,iBAAyB,EACzB,mBAA2B,CAAC;IAE5B,MAAM,aAAa,GAAG,iBAAiB,GAAG,gBAAgB,CAAC;IAE3D,+BAA+B;IAC/B,IAAI,iBAAiB,IAAI,GAAG,EAAE,CAAC;QAC7B,IAAI,aAAa,GAAG,IAAI,IAAI,cAAc,GAAG,iBAAiB,GAAG,CAAC,EAAE,CAAC;YACnE,OAAO,WAAW,CAAC;QACrB,CAAC;QACD,OAAO,UAAU,CAAC;IACpB,CAAC;IAED,iCAAiC;IACjC,IAAI,iBAAiB,IAAI,GAAG,EAAE,CAAC;QAC7B,IAAI,aAAa,GAAG,CAAC,EAAE,CAAC;YACtB,OAAO,UAAU,CAAC;QACpB,CAAC;QACD,OAAO,aAAa,CAAC;IACvB,CAAC;IAED,6BAA6B;IAC7B,IAAI,cAAc,GAAG,iBAAiB,EAAE,CAAC;QACvC,OAAO,UAAU,CAAC;IACpB,CAAC;IACD,OAAO,WAAW,CAAC;AACrB,CAAC;AAYD,MAAM,UAAU,kBAAkB,CAChC,OAAsC,EACtC,QAAgB,EAChB,MAA0B;IAE1B,MAAM,KAAK,GAAG,KAAK,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;IAC7C,MAAM,MAAM,GAAG,MAAM,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,EAAE,CAAC;IAC3D,MAAM,MAAM,GAAG,MAAM,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC;IAC1D,MAAM,IAAI,GAAG,MAAM,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,IAAI,GAAG,IAAI,CAAC,CAAC,KAAK,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC;IAE1E,IAAI,CAAC,OAAO,KAAK,MAAM,IAAI,OAAO,KAAK,WAAW,CAAC,IAAI,KAAK,IAAI,GAAG,EAAE,CAAC;QACpE,MAAM,OAAO,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC;YAC/B,CAAC,CAAC,kBAAkB,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG;YAChE,CAAC,CAAC,EAAE,CAAC;QACP,OAAO;YACL,WAAW,EAAE,GAAG,OAAO,WAAW,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,kCAAkC,OAAO,EAAE;SAC9F,CAAC;IACJ,CAAC;IAED,IAAI,OAAO,KAAK,MAAM,IAAI,OAAO,KAAK,WAAW,EAAE,CAAC;QAClD,MAAM,UAAU,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC;YAClC,CAAC,CAAC,qBAAqB,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG;YACnE,CAAC,CAAC,EAAE,CAAC;QACP,MAAM,eAAe,GAAG;YACtB,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,YAAY,CAAC,CAAC,SAAS,aAAa,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC;YAC/E,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,eAAe,CAAC,CAAC,SAAS,aAAa,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC;SACjF,CAAC;QACF,IAAI,eAAe,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACjC,eAAe,CAAC,IAAI,CAAC,2DAA2D,CAAC,CAAC;QACpF,CAAC;QACD,OAAO;YACL,WAAW,EAAE,GAAG,OAAO,WAAW,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,2CAA2C,UAAU,EAAE;YACzG,eAAe;SAChB,CAAC;IACJ,CAAC;IAED,IAAI,OAAO,KAAK,aAAa,EAAE,CAAC;QAC9B,IAAI,KAAK,GAAG,GAAG,EAAE,CAAC;YAChB,MAAM,WAAW,GAAG;gBAClB,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,CAAC,SAAS,oBAAoB,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;gBACpF,gDAAgD;gBAChD,8CAA8C;aAC/C,CAAC;YACF,OAAO;gBACL,WAAW,EAAE,GAAG,OAAO,WAAW,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,oDAAoD;gBACtG,eAAe,EAAE,WAAW;aAC7B,CAAC;QACJ,CAAC;QACD,MAAM,eAAe,GAAG;YACtB,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,CAAC,SAAS,oBAAoB,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;YACpF,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,eAAe,CAAC,CAAC,SAAS,aAAa,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC;SACjF,CAAC;QACF,IAAI,eAAe,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACjC,eAAe,CAAC,IAAI,CAAC,6CAA6C,CAAC,CAAC;QACtE,CAAC;QACD,OAAO;YACL,WAAW,EAAE,GAAG,OAAO,WAAW,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,oDAAoD;YACtG,eAAe;SAChB,CAAC;IACJ,CAAC;IAED,IAAI,OAAO,KAAK,UAAU,EAAE,CAAC;QAC3B,MAAM,eAAe,GAAG;YACtB,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,eAAe,CAAC,CAAC,SAAS,aAAa,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC;YAChF,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,YAAY,CAAC,CAAC,SAAS,aAAa,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC;SAChF,CAAC;QACF,IAAI,eAAe,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACjC,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC;gBACpC,CAAC,CAAC,+FAA+F;gBACjG,CAAC,CAAC,qDAAqD,CAAC,CAAC;QAC7D,CAAC;QACD,OAAO;YACL,WAAW,EAAE,GAAG,OAAO,WAAW,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,mCAAmC;YACrF,eAAe;SAChB,CAAC;IACJ,CAAC;IAED,IAAI,OAAO,KAAK,UAAU,EAAE,CAAC;QAC3B,MAAM,eAAe,GAAG;YACtB,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,YAAY,CAAC,CAAC,SAAS,aAAa,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC;YAC/E,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,eAAe,CAAC,CAAC,SAAS,aAAa,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC;SACjF,CAAC;QACF,IAAI,eAAe,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACjC,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC;gBACpC,CAAC,CAAC,+FAA+F;gBACjG,CAAC,CAAC,wDAAwD,CAAC,CAAC;QAChE,CAAC;QACD,OAAO;YACL,WAAW,EAAE,GAAG,OAAO,WAAW,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,oDAAoD;YACtG,eAAe;SAChB,CAAC;IACJ,CAAC;IAED,8DAA8D;IAC9D,MAAM,OAAO,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC;QAC/B,CAAC,CAAC,aAAa,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG;QAC3D,CAAC,CAAC,EAAE,CAAC;IACP,MAAM,QAAQ,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC;QAC9B,CAAC,CAAC,uBAAuB,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG;QACnE,CAAC,CAAC,EAAE,CAAC;IACP,OAAO;QACL,WAAW,EAAE,GAAG,OAAO,WAAW,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,oBAAoB,OAAO,GAAG,QAAQ,EAAE;KAC3F,CAAC;AACJ,CAAC;AAED,MAAM,cAAc,GAAgC;IAClD,SAAS,EAAE,oBAAoB;IAC/B,QAAQ,EAAE,sBAAsB;IAChC,QAAQ,EAAE,eAAe;IACzB,WAAW,EAAE,YAAY;IACzB,SAAS,EAAE,YAAY;CACxB,CAAC;AAEF,SAAS,aAAa,CAAC,CAAS;IAC9B,OAAO,MAAM,CAAC,MAAM,CAAC,cAAc,EAAE,CAAC,CAAC,CAAC;AAC1C,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,OAAe;IAC1C,OAAO,aAAa,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,cAAc,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC;AACpE,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,OAAoB;IAC/C,QAAQ,OAAO,EAAE,CAAC;QAChB,KAAK,WAAW;YACd,OAAO,OAAO,CAAC;QACjB,KAAK,UAAU;YACb,OAAO,QAAQ,CAAC;QAClB,KAAK,aAAa;YAChB,OAAO,QAAQ,CAAC;QAClB,KAAK,UAAU;YACb,OAAO,MAAM,CAAC;QAChB,KAAK,WAAW;YACd,OAAO,KAAK,CAAC;IACjB,CAAC;AACH,CAAC"}
|