audrey 0.23.1 → 1.0.1

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.
Files changed (250) hide show
  1. package/CHANGELOG.md +101 -15
  2. package/LICENSE +21 -21
  3. package/README.md +232 -6
  4. package/SECURITY.md +2 -1
  5. package/benchmarks/adapter-kit.mjs +20 -0
  6. package/benchmarks/adapter-self-test.mjs +166 -0
  7. package/benchmarks/adapters/example-allow.mjs +28 -0
  8. package/benchmarks/adapters/mem0-platform.mjs +267 -0
  9. package/benchmarks/adapters/registry.json +51 -0
  10. package/benchmarks/adapters/zep-cloud.mjs +280 -0
  11. package/benchmarks/baselines.js +169 -0
  12. package/benchmarks/build-leaderboard.mjs +170 -0
  13. package/benchmarks/cases.js +537 -0
  14. package/benchmarks/create-conformance-card.mjs +139 -0
  15. package/benchmarks/create-submission-bundle.mjs +176 -0
  16. package/benchmarks/dry-run-external-adapters.mjs +165 -0
  17. package/benchmarks/guardbench.js +1125 -0
  18. package/benchmarks/output/adapter-self-test/guardbench-adapter-self-test.json +50 -0
  19. package/benchmarks/output/external/guardbench-external-dry-run.json +69 -0
  20. package/benchmarks/output/external/guardbench-external-evidence.json +56 -0
  21. package/benchmarks/output/guardbench-conformance-card.json +63 -0
  22. package/benchmarks/output/guardbench-manifest.json +414 -0
  23. package/benchmarks/output/guardbench-raw.json +1271 -0
  24. package/benchmarks/output/guardbench-summary.json +2107 -0
  25. package/benchmarks/output/leaderboard/guardbench-leaderboard.json +93 -0
  26. package/benchmarks/output/leaderboard/guardbench-leaderboard.md +7 -0
  27. package/benchmarks/output/submission-bundle/guardbench-conformance-card.json +63 -0
  28. package/benchmarks/output/submission-bundle/guardbench-manifest.json +414 -0
  29. package/benchmarks/output/submission-bundle/guardbench-raw.json +1271 -0
  30. package/benchmarks/output/submission-bundle/guardbench-summary.json +2107 -0
  31. package/benchmarks/output/submission-bundle/schemas/guardbench-adapter-registry.schema.json +69 -0
  32. package/benchmarks/output/submission-bundle/schemas/guardbench-adapter-self-test.schema.json +156 -0
  33. package/benchmarks/output/submission-bundle/schemas/guardbench-conformance-card.schema.json +184 -0
  34. package/benchmarks/output/submission-bundle/schemas/guardbench-external-dry-run.schema.json +74 -0
  35. package/benchmarks/output/submission-bundle/schemas/guardbench-external-evidence.schema.json +108 -0
  36. package/benchmarks/output/submission-bundle/schemas/guardbench-external-run.schema.json +160 -0
  37. package/benchmarks/output/submission-bundle/schemas/guardbench-leaderboard.schema.json +179 -0
  38. package/benchmarks/output/submission-bundle/schemas/guardbench-manifest.schema.json +213 -0
  39. package/benchmarks/output/submission-bundle/schemas/guardbench-publication-verification.schema.json +47 -0
  40. package/benchmarks/output/submission-bundle/schemas/guardbench-raw.schema.json +184 -0
  41. package/benchmarks/output/submission-bundle/schemas/guardbench-submission-manifest.schema.json +151 -0
  42. package/benchmarks/output/submission-bundle/schemas/guardbench-summary.schema.json +249 -0
  43. package/benchmarks/output/submission-bundle/submission-manifest.json +131 -0
  44. package/benchmarks/output/submission-bundle/validation-report.json +31 -0
  45. package/benchmarks/output/summary.json +2354 -0
  46. package/benchmarks/perf-snapshot.js +304 -0
  47. package/benchmarks/perf.bench.js +161 -0
  48. package/benchmarks/public-paths.mjs +78 -0
  49. package/benchmarks/reference-results.js +70 -0
  50. package/benchmarks/report.js +259 -0
  51. package/benchmarks/run-external-guardbench.mjs +281 -0
  52. package/benchmarks/run.js +682 -0
  53. package/benchmarks/schemas/guardbench-adapter-registry.schema.json +69 -0
  54. package/benchmarks/schemas/guardbench-adapter-self-test.schema.json +156 -0
  55. package/benchmarks/schemas/guardbench-conformance-card.schema.json +184 -0
  56. package/benchmarks/schemas/guardbench-external-dry-run.schema.json +74 -0
  57. package/benchmarks/schemas/guardbench-external-evidence.schema.json +108 -0
  58. package/benchmarks/schemas/guardbench-external-run.schema.json +160 -0
  59. package/benchmarks/schemas/guardbench-leaderboard.schema.json +179 -0
  60. package/benchmarks/schemas/guardbench-manifest.schema.json +213 -0
  61. package/benchmarks/schemas/guardbench-publication-verification.schema.json +47 -0
  62. package/benchmarks/schemas/guardbench-raw.schema.json +184 -0
  63. package/benchmarks/schemas/guardbench-submission-manifest.schema.json +151 -0
  64. package/benchmarks/schemas/guardbench-summary.schema.json +249 -0
  65. package/benchmarks/snapshots/perf-0.22.2.json +123 -0
  66. package/benchmarks/snapshots/perf-0.23.0.json +123 -0
  67. package/benchmarks/validate-adapter-module.mjs +104 -0
  68. package/benchmarks/validate-adapter-registry.mjs +134 -0
  69. package/benchmarks/validate-adapter-self-test.mjs +96 -0
  70. package/benchmarks/validate-guardbench-artifacts.mjs +343 -0
  71. package/benchmarks/verify-external-evidence.mjs +296 -0
  72. package/benchmarks/verify-publication-artifacts.mjs +286 -0
  73. package/benchmarks/verify-submission-bundle.mjs +167 -0
  74. package/dist/mcp-server/config.d.ts +1 -1
  75. package/dist/mcp-server/config.d.ts.map +1 -1
  76. package/dist/mcp-server/config.js +1 -1
  77. package/dist/mcp-server/config.js.map +1 -1
  78. package/dist/mcp-server/index.d.ts +65 -3
  79. package/dist/mcp-server/index.d.ts.map +1 -1
  80. package/dist/mcp-server/index.js +675 -157
  81. package/dist/mcp-server/index.js.map +1 -1
  82. package/dist/src/action-key.d.ts +9 -0
  83. package/dist/src/action-key.d.ts.map +1 -0
  84. package/dist/src/action-key.js +49 -0
  85. package/dist/src/action-key.js.map +1 -0
  86. package/dist/src/adaptive.js +5 -5
  87. package/dist/src/affect.js +8 -8
  88. package/dist/src/audrey.d.ts +13 -0
  89. package/dist/src/audrey.d.ts.map +1 -1
  90. package/dist/src/audrey.js +68 -3
  91. package/dist/src/audrey.js.map +1 -1
  92. package/dist/src/capsule.js +4 -4
  93. package/dist/src/causal.js +3 -3
  94. package/dist/src/consolidate.js +48 -48
  95. package/dist/src/controller.d.ts +78 -6
  96. package/dist/src/controller.d.ts.map +1 -1
  97. package/dist/src/controller.js +273 -53
  98. package/dist/src/controller.js.map +1 -1
  99. package/dist/src/db.js +172 -172
  100. package/dist/src/decay.js +8 -8
  101. package/dist/src/embedding.d.ts +2 -1
  102. package/dist/src/embedding.d.ts.map +1 -1
  103. package/dist/src/embedding.js +39 -29
  104. package/dist/src/embedding.js.map +1 -1
  105. package/dist/src/encode.js +6 -6
  106. package/dist/src/feedback.d.ts +6 -0
  107. package/dist/src/feedback.d.ts.map +1 -1
  108. package/dist/src/feedback.js +6 -0
  109. package/dist/src/feedback.js.map +1 -1
  110. package/dist/src/forget.js +12 -12
  111. package/dist/src/hybrid-recall.js +9 -9
  112. package/dist/src/impact.js +6 -6
  113. package/dist/src/import.d.ts +3 -3
  114. package/dist/src/import.js +41 -41
  115. package/dist/src/index.d.ts +5 -4
  116. package/dist/src/index.d.ts.map +1 -1
  117. package/dist/src/index.js +3 -3
  118. package/dist/src/index.js.map +1 -1
  119. package/dist/src/interference.js +14 -14
  120. package/dist/src/introspect.js +18 -18
  121. package/dist/src/preflight.d.ts.map +1 -1
  122. package/dist/src/preflight.js +41 -0
  123. package/dist/src/preflight.js.map +1 -1
  124. package/dist/src/promote.js +7 -7
  125. package/dist/src/prompts.js +118 -118
  126. package/dist/src/recall.js +30 -30
  127. package/dist/src/reflexes.d.ts +1 -0
  128. package/dist/src/reflexes.d.ts.map +1 -1
  129. package/dist/src/reflexes.js +3 -0
  130. package/dist/src/reflexes.js.map +1 -1
  131. package/dist/src/rollback.js +4 -4
  132. package/dist/src/routes.d.ts.map +1 -1
  133. package/dist/src/routes.js +71 -2
  134. package/dist/src/routes.js.map +1 -1
  135. package/dist/src/validate.js +25 -25
  136. package/docs/AUDREY_PAPER_OUTLINE.md +175 -0
  137. package/docs/MEMORY_BENCHMARKING.md +59 -0
  138. package/docs/PRODUCTION_BACKLOG.md +304 -0
  139. package/docs/paper/00-master.md +48 -0
  140. package/docs/paper/01-introduction.md +27 -0
  141. package/docs/paper/02-related-work.md +47 -0
  142. package/docs/paper/03-problem-definition.md +108 -0
  143. package/docs/paper/04-design.md +164 -0
  144. package/docs/paper/05-guardbench-spec.md +412 -0
  145. package/docs/paper/06-implementation.md +113 -0
  146. package/docs/paper/07-evaluation.md +168 -0
  147. package/docs/paper/08-discussion-limitations.md +61 -0
  148. package/docs/paper/09-conclusion.md +11 -0
  149. package/docs/paper/SUBMISSION_README.md +162 -0
  150. package/docs/paper/appendix-a-demo-transcript.md +114 -0
  151. package/docs/paper/arxiv-compile-report.schema.json +116 -0
  152. package/docs/paper/arxiv-source.schema.json +61 -0
  153. package/docs/paper/audrey-paper-v1.md +1106 -0
  154. package/docs/paper/browser-launch-plan.json +209 -0
  155. package/docs/paper/browser-launch-plan.schema.json +100 -0
  156. package/docs/paper/browser-launch-results.json +86 -0
  157. package/docs/paper/browser-launch-results.schema.json +66 -0
  158. package/docs/paper/claim-register.json +138 -0
  159. package/docs/paper/claim-register.schema.json +81 -0
  160. package/docs/paper/evidence-ledger.md +103 -0
  161. package/docs/paper/output/arxiv/README-arxiv.txt +8 -0
  162. package/docs/paper/output/arxiv/arxiv-manifest.json +41 -0
  163. package/docs/paper/output/arxiv/main.tex +949 -0
  164. package/docs/paper/output/arxiv/references.bib +222 -0
  165. package/docs/paper/output/arxiv-compile-report.json +24 -0
  166. package/docs/paper/output/submission-bundle/LICENSE +21 -0
  167. package/docs/paper/output/submission-bundle/README.md +555 -0
  168. package/docs/paper/output/submission-bundle/benchmarks/output/adapter-self-test/guardbench-adapter-self-test.json +50 -0
  169. package/docs/paper/output/submission-bundle/benchmarks/output/external/guardbench-external-dry-run.json +69 -0
  170. package/docs/paper/output/submission-bundle/benchmarks/output/external/guardbench-external-evidence.json +56 -0
  171. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-conformance-card.json +63 -0
  172. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-manifest.json +414 -0
  173. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-raw.json +1271 -0
  174. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-summary.json +2107 -0
  175. package/docs/paper/output/submission-bundle/benchmarks/output/leaderboard/guardbench-leaderboard.json +93 -0
  176. package/docs/paper/output/submission-bundle/benchmarks/output/leaderboard/guardbench-leaderboard.md +7 -0
  177. package/docs/paper/output/submission-bundle/benchmarks/output/submission-bundle/submission-manifest.json +131 -0
  178. package/docs/paper/output/submission-bundle/benchmarks/output/submission-bundle/validation-report.json +31 -0
  179. package/docs/paper/output/submission-bundle/benchmarks/output/summary.json +2354 -0
  180. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-adapter-registry.schema.json +69 -0
  181. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-adapter-self-test.schema.json +156 -0
  182. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-conformance-card.schema.json +184 -0
  183. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-external-dry-run.schema.json +74 -0
  184. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-external-evidence.schema.json +108 -0
  185. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-external-run.schema.json +160 -0
  186. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-leaderboard.schema.json +179 -0
  187. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-manifest.schema.json +213 -0
  188. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-publication-verification.schema.json +47 -0
  189. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-raw.schema.json +184 -0
  190. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-submission-manifest.schema.json +151 -0
  191. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-summary.schema.json +249 -0
  192. package/docs/paper/output/submission-bundle/docs/AUDREY_PAPER_OUTLINE.md +175 -0
  193. package/docs/paper/output/submission-bundle/docs/paper/00-master.md +48 -0
  194. package/docs/paper/output/submission-bundle/docs/paper/01-introduction.md +27 -0
  195. package/docs/paper/output/submission-bundle/docs/paper/02-related-work.md +47 -0
  196. package/docs/paper/output/submission-bundle/docs/paper/03-problem-definition.md +108 -0
  197. package/docs/paper/output/submission-bundle/docs/paper/04-design.md +164 -0
  198. package/docs/paper/output/submission-bundle/docs/paper/05-guardbench-spec.md +412 -0
  199. package/docs/paper/output/submission-bundle/docs/paper/06-implementation.md +113 -0
  200. package/docs/paper/output/submission-bundle/docs/paper/07-evaluation.md +168 -0
  201. package/docs/paper/output/submission-bundle/docs/paper/08-discussion-limitations.md +61 -0
  202. package/docs/paper/output/submission-bundle/docs/paper/09-conclusion.md +11 -0
  203. package/docs/paper/output/submission-bundle/docs/paper/SUBMISSION_README.md +162 -0
  204. package/docs/paper/output/submission-bundle/docs/paper/appendix-a-demo-transcript.md +114 -0
  205. package/docs/paper/output/submission-bundle/docs/paper/arxiv-compile-report.schema.json +116 -0
  206. package/docs/paper/output/submission-bundle/docs/paper/arxiv-source.schema.json +61 -0
  207. package/docs/paper/output/submission-bundle/docs/paper/audrey-paper-v1.md +1106 -0
  208. package/docs/paper/output/submission-bundle/docs/paper/browser-launch-plan.json +209 -0
  209. package/docs/paper/output/submission-bundle/docs/paper/browser-launch-plan.schema.json +100 -0
  210. package/docs/paper/output/submission-bundle/docs/paper/browser-launch-results.json +86 -0
  211. package/docs/paper/output/submission-bundle/docs/paper/browser-launch-results.schema.json +66 -0
  212. package/docs/paper/output/submission-bundle/docs/paper/claim-register.json +138 -0
  213. package/docs/paper/output/submission-bundle/docs/paper/claim-register.schema.json +81 -0
  214. package/docs/paper/output/submission-bundle/docs/paper/evidence-ledger.md +103 -0
  215. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/README-arxiv.txt +8 -0
  216. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/arxiv-manifest.json +41 -0
  217. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/main.tex +949 -0
  218. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/references.bib +222 -0
  219. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv-compile-report.json +24 -0
  220. package/docs/paper/output/submission-bundle/docs/paper/paper-submission-bundle.schema.json +70 -0
  221. package/docs/paper/output/submission-bundle/docs/paper/publication-pack.json +81 -0
  222. package/docs/paper/output/submission-bundle/docs/paper/publication-pack.schema.json +60 -0
  223. package/docs/paper/output/submission-bundle/docs/paper/references.bib +222 -0
  224. package/docs/paper/output/submission-bundle/package.json +212 -0
  225. package/docs/paper/output/submission-bundle/paper-submission-manifest.json +379 -0
  226. package/docs/paper/paper-submission-bundle.schema.json +70 -0
  227. package/docs/paper/publication-pack.json +81 -0
  228. package/docs/paper/publication-pack.schema.json +60 -0
  229. package/docs/paper/references.bib +222 -0
  230. package/package.json +87 -4
  231. package/scripts/audit-release-completion.mjs +362 -0
  232. package/scripts/create-arxiv-source.mjs +362 -0
  233. package/scripts/create-paper-submission-bundle.mjs +210 -0
  234. package/scripts/finalize-release.mjs +526 -0
  235. package/scripts/prepare-release-cut.mjs +269 -0
  236. package/scripts/publish-release-bundle.mjs +209 -0
  237. package/scripts/publish-release-github-api.mjs +429 -0
  238. package/scripts/run-vitest.mjs +34 -0
  239. package/scripts/smoke-cli.js +92 -0
  240. package/scripts/sync-paper-artifacts.mjs +109 -0
  241. package/scripts/verify-arxiv-compile.mjs +440 -0
  242. package/scripts/verify-arxiv-source.mjs +194 -0
  243. package/scripts/verify-browser-launch-plan.mjs +237 -0
  244. package/scripts/verify-browser-launch-results.mjs +285 -0
  245. package/scripts/verify-paper-artifacts.mjs +338 -0
  246. package/scripts/verify-paper-claims.mjs +226 -0
  247. package/scripts/verify-paper-submission-bundle.mjs +207 -0
  248. package/scripts/verify-publication-pack.mjs +196 -0
  249. package/scripts/verify-python-package.py +201 -0
  250. package/scripts/verify-release-readiness.mjs +785 -0
@@ -0,0 +1,176 @@
1
+ import { cpSync, existsSync, mkdirSync, readFileSync, rmSync, writeFileSync } from 'node:fs';
2
+ import { createHash } from 'node:crypto';
3
+ import { basename, isAbsolute, join, relative, resolve } from 'node:path';
4
+ import { writeGuardBenchConformanceCard } from './create-conformance-card.mjs';
5
+ import { validateGuardBenchArtifacts } from './validate-guardbench-artifacts.mjs';
6
+ import { publicPath } from './public-paths.mjs';
7
+
8
+ const REQUIRED_ARTIFACTS = [
9
+ 'guardbench-manifest.json',
10
+ 'guardbench-summary.json',
11
+ 'guardbench-raw.json',
12
+ 'guardbench-conformance-card.json',
13
+ ];
14
+ const OPTIONAL_ARTIFACTS = ['external-run-metadata.json'];
15
+ const SCHEMA_FILES = [
16
+ 'guardbench-adapter-registry.schema.json',
17
+ 'guardbench-adapter-self-test.schema.json',
18
+ 'guardbench-external-dry-run.schema.json',
19
+ 'guardbench-external-evidence.schema.json',
20
+ 'guardbench-publication-verification.schema.json',
21
+ 'guardbench-manifest.schema.json',
22
+ 'guardbench-summary.schema.json',
23
+ 'guardbench-raw.schema.json',
24
+ 'guardbench-external-run.schema.json',
25
+ 'guardbench-conformance-card.schema.json',
26
+ 'guardbench-leaderboard.schema.json',
27
+ 'guardbench-submission-manifest.schema.json',
28
+ ];
29
+
30
+ function readJson(path) {
31
+ return JSON.parse(readFileSync(path, 'utf-8'));
32
+ }
33
+
34
+ function sha256File(path) {
35
+ return createHash('sha256').update(readFileSync(path)).digest('hex');
36
+ }
37
+
38
+ function copyFileInto(sourceDir, outDir, file) {
39
+ const source = join(sourceDir, file);
40
+ if (!existsSync(source)) return null;
41
+ const target = join(outDir, file);
42
+ cpSync(source, target);
43
+ return target;
44
+ }
45
+
46
+ export function bundleRelativeFilePath(path, root) {
47
+ const relativePath = relative(root, path);
48
+ if (!relativePath || relativePath.startsWith('..') || isAbsolute(relativePath)) {
49
+ throw new Error(`Cannot add file outside GuardBench submission bundle: ${path}`);
50
+ }
51
+ return relativePath.replaceAll('\\', '/');
52
+ }
53
+
54
+ function fileRecord(path, root) {
55
+ return {
56
+ path: bundleRelativeFilePath(path, root),
57
+ bytes: readFileSync(path).byteLength,
58
+ sha256: sha256File(path),
59
+ };
60
+ }
61
+
62
+ export function writeGuardBenchSubmissionBundle(options = {}) {
63
+ const sourceDir = resolve(options.dir ?? 'benchmarks/output');
64
+ const outDir = resolve(options.outDir ?? join(sourceDir, 'submission-bundle'));
65
+ const schemasDir = resolve(options.schemasDir ?? 'benchmarks/schemas');
66
+
67
+ writeGuardBenchConformanceCard({ dir: sourceDir });
68
+ const sourceValidation = validateGuardBenchArtifacts({ dir: sourceDir, schemasDir });
69
+ if (!sourceValidation.ok) {
70
+ throw new Error(`Cannot create GuardBench submission bundle from invalid artifacts: ${sourceValidation.failures.join('; ')}`);
71
+ }
72
+
73
+ rmSync(outDir, { recursive: true, force: true });
74
+ mkdirSync(join(outDir, 'schemas'), { recursive: true });
75
+
76
+ const copied = [];
77
+ for (const file of [...REQUIRED_ARTIFACTS, ...OPTIONAL_ARTIFACTS]) {
78
+ const target = copyFileInto(sourceDir, outDir, file);
79
+ if (target) copied.push(target);
80
+ }
81
+ for (const file of SCHEMA_FILES) {
82
+ const target = join(outDir, 'schemas', file);
83
+ cpSync(join(schemasDir, file), target);
84
+ copied.push(target);
85
+ }
86
+
87
+ const bundleValidation = validateGuardBenchArtifacts({
88
+ dir: outDir,
89
+ schemasDir: join(outDir, 'schemas'),
90
+ });
91
+ const validationReportPath = join(outDir, 'validation-report.json');
92
+ writeFileSync(validationReportPath, `${JSON.stringify({
93
+ generatedAt: new Date().toISOString(),
94
+ sourceValidation,
95
+ bundleValidation,
96
+ }, null, 2)}\n`, 'utf-8');
97
+ copied.push(validationReportPath);
98
+
99
+ const card = readJson(join(outDir, 'guardbench-conformance-card.json'));
100
+ const manifestPath = join(outDir, 'submission-manifest.json');
101
+ const manifest = {
102
+ schemaVersion: '1.0.0',
103
+ suite: 'GuardBench submission bundle',
104
+ generatedAt: new Date().toISOString(),
105
+ sourceDir: publicPath(sourceDir),
106
+ subject: card.subject,
107
+ score: card.score,
108
+ conformance: card.conformance,
109
+ validation: bundleValidation,
110
+ files: copied.map(path => fileRecord(path, outDir)).sort((a, b) => a.path.localeCompare(b.path)),
111
+ };
112
+ writeFileSync(manifestPath, `${JSON.stringify(manifest, null, 2)}\n`, 'utf-8');
113
+
114
+ const finalFiles = copied.map(path => fileRecord(path, outDir)).sort((a, b) => a.path.localeCompare(b.path));
115
+ manifest.files = finalFiles;
116
+ writeFileSync(manifestPath, `${JSON.stringify(manifest, null, 2)}\n`, 'utf-8');
117
+
118
+ return {
119
+ outDir,
120
+ manifestPath,
121
+ validation: bundleValidation,
122
+ subject: card.subject,
123
+ files: finalFiles.map(record => record.path),
124
+ };
125
+ }
126
+
127
+ function parseArgs(argv = process.argv.slice(2)) {
128
+ const args = {
129
+ dir: 'benchmarks/output',
130
+ outDir: null,
131
+ schemasDir: 'benchmarks/schemas',
132
+ json: false,
133
+ };
134
+ for (let i = 0; i < argv.length; i++) {
135
+ const token = argv[i];
136
+ if ((token === '--dir' || token === '--source-dir') && argv[i + 1]) args.dir = argv[++i];
137
+ else if (token === '--out-dir' && argv[i + 1]) args.outDir = argv[++i];
138
+ else if (token === '--schemas-dir' && argv[i + 1]) args.schemasDir = argv[++i];
139
+ else if (token === '--json') args.json = true;
140
+ else if (token === '--help' || token === '-h') args.help = true;
141
+ else throw new Error(`Unknown argument: ${token}`);
142
+ }
143
+ return args;
144
+ }
145
+
146
+ function usage() {
147
+ return [
148
+ 'Usage: node benchmarks/create-submission-bundle.mjs [--dir benchmarks/output] [--out-dir <dir>] [--json]',
149
+ '',
150
+ 'Creates a portable GuardBench submission bundle containing artifacts, schemas,',
151
+ 'a conformance card, validation report, and submission manifest.',
152
+ ].join('\n');
153
+ }
154
+
155
+ async function main() {
156
+ const args = parseArgs();
157
+ if (args.help) {
158
+ console.log(usage());
159
+ return;
160
+ }
161
+ const result = writeGuardBenchSubmissionBundle(args);
162
+ if (args.json) console.log(JSON.stringify(result, null, 2));
163
+ else {
164
+ console.log(`GuardBench submission bundle: ${result.outDir}`);
165
+ console.log(`Subject: ${result.subject.name}`);
166
+ console.log(`Validation: ${result.validation.ok ? 'passed' : 'failed'}`);
167
+ }
168
+ if (!result.validation.ok) process.exit(1);
169
+ }
170
+
171
+ if (process.argv[1] && resolve(process.argv[1]).endsWith(basename(import.meta.url))) {
172
+ main().catch(error => {
173
+ console.error(error.stack ?? error.message);
174
+ process.exit(1);
175
+ });
176
+ }
@@ -0,0 +1,165 @@
1
+ import { mkdirSync, readFileSync, writeFileSync } from 'node:fs';
2
+ import { dirname, join, resolve } from 'node:path';
3
+ import { fileURLToPath } from 'node:url';
4
+ import { buildExternalGuardBenchRun, writeExternalRunMetadata } from './run-external-guardbench.mjs';
5
+ import { validateAdapterRegistry } from './validate-adapter-registry.mjs';
6
+ import { validateSchema } from './validate-guardbench-artifacts.mjs';
7
+ import { publicCommand, publicPath } from './public-paths.mjs';
8
+
9
+ const DEFAULT_REGISTRY = 'benchmarks/adapters/registry.json';
10
+ const DEFAULT_OUT_ROOT = 'benchmarks/output/external';
11
+ const DEFAULT_OUT = 'benchmarks/output/external/guardbench-external-dry-run.json';
12
+ const DEFAULT_SCHEMA = 'benchmarks/schemas/guardbench-external-dry-run.schema.json';
13
+
14
+ export function parseExternalDryRunArgs(argv = process.argv.slice(2)) {
15
+ const args = {
16
+ registry: DEFAULT_REGISTRY,
17
+ outRoot: DEFAULT_OUT_ROOT,
18
+ out: DEFAULT_OUT,
19
+ includeCredentialFree: false,
20
+ json: false,
21
+ noWrite: false,
22
+ };
23
+
24
+ for (let i = 0; i < argv.length; i++) {
25
+ const token = argv[i];
26
+ if (token === '--registry' && argv[i + 1]) args.registry = argv[++i];
27
+ else if (token === '--out-root' && argv[i + 1]) args.outRoot = argv[++i];
28
+ else if (token === '--out' && argv[i + 1]) args.out = argv[++i];
29
+ else if (token === '--include-credential-free') args.includeCredentialFree = true;
30
+ else if (token === '--json') args.json = true;
31
+ else if (token === '--no-write') args.noWrite = true;
32
+ else if (token === '--help' || token === '-h') args.help = true;
33
+ else throw new Error(`Unknown argument: ${token}`);
34
+ }
35
+
36
+ return args;
37
+ }
38
+
39
+ function usage() {
40
+ return `Usage: node benchmarks/dry-run-external-adapters.mjs [options]
41
+
42
+ Options:
43
+ --registry <path> Adapter registry JSON. Default: ${DEFAULT_REGISTRY}.
44
+ --out-root <path> Root directory for dry-run metadata. Default: ${DEFAULT_OUT_ROOT}.
45
+ --out <path> Matrix JSON report path. Default: ${DEFAULT_OUT}.
46
+ --include-credential-free Include credential-free registry adapters.
47
+ --json Print the machine-readable dry-run matrix.
48
+ --no-write Do not write the matrix JSON report.
49
+ `;
50
+ }
51
+
52
+ function readJson(path) {
53
+ return JSON.parse(readFileSync(path, 'utf-8'));
54
+ }
55
+
56
+ export function validateExternalAdapterDryRunMatrix(matrix, options = {}) {
57
+ const schema = options.schemaObject ?? readJson(options.schema ?? DEFAULT_SCHEMA);
58
+ return validateSchema(matrix, schema, 'guardbench-external-dry-run');
59
+ }
60
+
61
+ export function writeExternalAdapterDryRunMatrix(matrix, out = DEFAULT_OUT) {
62
+ const outPath = resolve(out);
63
+ mkdirSync(dirname(outPath), { recursive: true });
64
+ writeFileSync(outPath, `${JSON.stringify(matrix, null, 2)}\n`, 'utf-8');
65
+ return outPath;
66
+ }
67
+
68
+ export async function buildExternalAdapterDryRunMatrix(options = {}) {
69
+ const registryPath = options.registry ?? DEFAULT_REGISTRY;
70
+ const outRoot = resolve(options.outRoot ?? DEFAULT_OUT_ROOT);
71
+ const registryValidation = await validateAdapterRegistry({ registry: registryPath });
72
+ const registry = readJson(registryPath);
73
+ const adapters = registry.adapters.filter(adapter =>
74
+ adapter.status === 'external-system'
75
+ && (options.includeCredentialFree || adapter.credentialMode === 'runtime-env'));
76
+ const rows = [];
77
+ const failures = [];
78
+
79
+ if (!registryValidation.ok) {
80
+ failures.push(...registryValidation.failures.map(failure => `registry: ${failure}`));
81
+ }
82
+
83
+ for (const adapter of adapters) {
84
+ const run = buildExternalGuardBenchRun({
85
+ adapter: adapter.id,
86
+ outDir: join(outRoot, adapter.id),
87
+ check: true,
88
+ json: true,
89
+ }, options.env ?? process.env);
90
+ const metadata = {
91
+ suite: 'GuardBench external adapter run',
92
+ startedAt: new Date().toISOString(),
93
+ adapter: run.adapter,
94
+ adapterPath: run.adapterPath,
95
+ outDir: run.outDir,
96
+ requiredEnv: run.requiredEnv,
97
+ missingEnv: run.missingEnv,
98
+ command: run.command,
99
+ validationCommand: run.validationCommand,
100
+ dryRun: true,
101
+ status: run.missingEnv.length ? 'dry-run-missing-env' : 'dry-run-ready',
102
+ };
103
+ const metadataPath = writeExternalRunMetadata(run.outDir, metadata);
104
+ rows.push({
105
+ id: adapter.id,
106
+ name: adapter.name,
107
+ credentialMode: adapter.credentialMode,
108
+ requiredEnv: run.requiredEnv,
109
+ missingEnv: run.missingEnv,
110
+ status: metadata.status,
111
+ command: publicCommand(run.command),
112
+ validationCommand: publicCommand(run.validationCommand),
113
+ metadataPath: publicPath(metadataPath),
114
+ });
115
+ }
116
+
117
+ const matrix = {
118
+ schemaVersion: '1.0.0',
119
+ suite: 'GuardBench external adapter dry-run matrix',
120
+ generatedAt: new Date().toISOString(),
121
+ ok: failures.length === 0,
122
+ registry: publicPath(resolve(registryPath)),
123
+ outRoot: publicPath(outRoot),
124
+ adapters: rows,
125
+ failures,
126
+ };
127
+ const schemaFailures = validateExternalAdapterDryRunMatrix(matrix, options);
128
+ if (schemaFailures.length > 0) {
129
+ throw new Error(`GuardBench external adapter dry-run schema validation failed: ${schemaFailures.join('; ')}`);
130
+ }
131
+ return matrix;
132
+ }
133
+
134
+ async function main() {
135
+ const args = parseExternalDryRunArgs();
136
+ if (args.help) {
137
+ console.log(usage());
138
+ return;
139
+ }
140
+
141
+ const matrix = await buildExternalAdapterDryRunMatrix(args);
142
+ const outPath = args.noWrite ? null : writeExternalAdapterDryRunMatrix(matrix, args.out);
143
+ if (args.json) {
144
+ console.log(JSON.stringify(matrix, null, 2));
145
+ } else if (matrix.ok) {
146
+ console.log(`GuardBench external adapter dry-run matrix passed: ${matrix.adapters.length} adapter(s)`);
147
+ for (const row of matrix.adapters) {
148
+ const missing = row.missingEnv.length ? `missing ${row.missingEnv.join(', ')}` : 'ready';
149
+ console.log(`- ${row.id}: ${missing}; metadata ${row.metadataPath}`);
150
+ }
151
+ if (outPath) console.log(`Matrix report: ${outPath}`);
152
+ } else {
153
+ console.error('GuardBench external adapter dry-run matrix failed:');
154
+ for (const failure of matrix.failures) console.error(`- ${failure}`);
155
+ }
156
+
157
+ if (!matrix.ok) process.exit(1);
158
+ }
159
+
160
+ if (process.argv[1] && resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
161
+ main().catch(error => {
162
+ console.error(error.stack ?? error.message);
163
+ process.exit(1);
164
+ });
165
+ }