audrey 1.1.3 → 1.2.0

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 (285) hide show
  1. package/CHANGELOG.md +93 -0
  2. package/README.md +39 -12
  3. package/dist/mcp-server/codex-hooks-probe.d.ts +22 -0
  4. package/dist/mcp-server/codex-hooks-probe.d.ts.map +1 -0
  5. package/dist/mcp-server/codex-hooks-probe.js +160 -0
  6. package/dist/mcp-server/codex-hooks-probe.js.map +1 -0
  7. package/dist/mcp-server/config.d.ts +26 -2
  8. package/dist/mcp-server/config.d.ts.map +1 -1
  9. package/dist/mcp-server/config.js +52 -20
  10. package/dist/mcp-server/config.js.map +1 -1
  11. package/dist/mcp-server/hooks.d.ts +28 -0
  12. package/dist/mcp-server/hooks.d.ts.map +1 -1
  13. package/dist/mcp-server/hooks.js +91 -2
  14. package/dist/mcp-server/hooks.js.map +1 -1
  15. package/dist/mcp-server/index.d.ts +242 -2
  16. package/dist/mcp-server/index.d.ts.map +1 -1
  17. package/dist/mcp-server/index.js +806 -240
  18. package/dist/mcp-server/index.js.map +1 -1
  19. package/dist/mcp-server/tool-schemas.d.ts +32 -2
  20. package/dist/mcp-server/tool-schemas.d.ts.map +1 -1
  21. package/dist/mcp-server/tool-schemas.js +55 -0
  22. package/dist/mcp-server/tool-schemas.js.map +1 -1
  23. package/dist/src/audrey.d.ts +14 -2
  24. package/dist/src/audrey.d.ts.map +1 -1
  25. package/dist/src/audrey.js +152 -28
  26. package/dist/src/audrey.js.map +1 -1
  27. package/dist/src/autopilot.d.ts +22 -0
  28. package/dist/src/autopilot.d.ts.map +1 -1
  29. package/dist/src/autopilot.js +271 -50
  30. package/dist/src/autopilot.js.map +1 -1
  31. package/dist/src/capsule.d.ts +23 -0
  32. package/dist/src/capsule.d.ts.map +1 -1
  33. package/dist/src/capsule.js +56 -11
  34. package/dist/src/capsule.js.map +1 -1
  35. package/dist/src/confidence.d.ts +9 -1
  36. package/dist/src/confidence.d.ts.map +1 -1
  37. package/dist/src/confidence.js +16 -2
  38. package/dist/src/confidence.js.map +1 -1
  39. package/dist/src/consolidate.d.ts +4 -0
  40. package/dist/src/consolidate.d.ts.map +1 -1
  41. package/dist/src/consolidate.js +206 -34
  42. package/dist/src/consolidate.js.map +1 -1
  43. package/dist/src/controller.d.ts +17 -1
  44. package/dist/src/controller.d.ts.map +1 -1
  45. package/dist/src/controller.js +244 -128
  46. package/dist/src/controller.js.map +1 -1
  47. package/dist/src/db.d.ts.map +1 -1
  48. package/dist/src/db.js +233 -56
  49. package/dist/src/db.js.map +1 -1
  50. package/dist/src/encode.d.ts +2 -1
  51. package/dist/src/encode.d.ts.map +1 -1
  52. package/dist/src/encode.js +41 -4
  53. package/dist/src/encode.js.map +1 -1
  54. package/dist/src/events.d.ts +48 -1
  55. package/dist/src/events.d.ts.map +1 -1
  56. package/dist/src/events.js +160 -62
  57. package/dist/src/events.js.map +1 -1
  58. package/dist/src/export.d.ts.map +1 -1
  59. package/dist/src/export.js +4 -4
  60. package/dist/src/export.js.map +1 -1
  61. package/dist/src/grounding.d.ts +111 -0
  62. package/dist/src/grounding.d.ts.map +1 -0
  63. package/dist/src/grounding.js +259 -0
  64. package/dist/src/grounding.js.map +1 -0
  65. package/dist/src/hybrid-recall.d.ts +16 -7
  66. package/dist/src/hybrid-recall.d.ts.map +1 -1
  67. package/dist/src/hybrid-recall.js +22 -11
  68. package/dist/src/hybrid-recall.js.map +1 -1
  69. package/dist/src/import.d.ts +10 -0
  70. package/dist/src/import.d.ts.map +1 -1
  71. package/dist/src/import.js +122 -22
  72. package/dist/src/import.js.map +1 -1
  73. package/dist/src/llm.d.ts.map +1 -1
  74. package/dist/src/llm.js +16 -0
  75. package/dist/src/llm.js.map +1 -1
  76. package/dist/src/preflight.d.ts.map +1 -1
  77. package/dist/src/preflight.js +89 -47
  78. package/dist/src/preflight.js.map +1 -1
  79. package/dist/src/project.d.ts +4 -1
  80. package/dist/src/project.d.ts.map +1 -1
  81. package/dist/src/project.js +65 -5
  82. package/dist/src/project.js.map +1 -1
  83. package/dist/src/recall.d.ts.map +1 -1
  84. package/dist/src/recall.js +138 -13
  85. package/dist/src/recall.js.map +1 -1
  86. package/dist/src/redact.d.ts +1 -1
  87. package/dist/src/redact.d.ts.map +1 -1
  88. package/dist/src/redact.js +185 -5
  89. package/dist/src/redact.js.map +1 -1
  90. package/dist/src/routes.d.ts.map +1 -1
  91. package/dist/src/routes.js +91 -18
  92. package/dist/src/routes.js.map +1 -1
  93. package/dist/src/trust.d.ts +35 -0
  94. package/dist/src/trust.d.ts.map +1 -0
  95. package/dist/src/trust.js +77 -0
  96. package/dist/src/trust.js.map +1 -0
  97. package/dist/src/types.d.ts +24 -0
  98. package/dist/src/types.d.ts.map +1 -1
  99. package/package.json +3 -37
  100. package/benchmarks/adapter-kit.mjs +0 -20
  101. package/benchmarks/adapter-self-test.mjs +0 -170
  102. package/benchmarks/adapters/example-allow.mjs +0 -31
  103. package/benchmarks/adapters/mem0-platform.mjs +0 -274
  104. package/benchmarks/adapters/registry.json +0 -51
  105. package/benchmarks/adapters/zep-cloud.mjs +0 -304
  106. package/benchmarks/baselines.js +0 -174
  107. package/benchmarks/build-leaderboard.mjs +0 -183
  108. package/benchmarks/cases.js +0 -549
  109. package/benchmarks/create-conformance-card.mjs +0 -148
  110. package/benchmarks/create-submission-bundle.mjs +0 -190
  111. package/benchmarks/dry-run-external-adapters.mjs +0 -177
  112. package/benchmarks/guardbench.js +0 -1265
  113. package/benchmarks/output/adapter-self-test/guardbench-adapter-self-test.json +0 -50
  114. package/benchmarks/output/external/guardbench-external-dry-run.json +0 -69
  115. package/benchmarks/output/external/guardbench-external-evidence.json +0 -56
  116. package/benchmarks/output/guardbench-conformance-card.json +0 -63
  117. package/benchmarks/output/guardbench-manifest.json +0 -414
  118. package/benchmarks/output/guardbench-raw.json +0 -1279
  119. package/benchmarks/output/guardbench-summary.json +0 -2123
  120. package/benchmarks/output/leaderboard/guardbench-leaderboard.json +0 -93
  121. package/benchmarks/output/leaderboard/guardbench-leaderboard.md +0 -7
  122. package/benchmarks/output/submission-bundle/guardbench-conformance-card.json +0 -63
  123. package/benchmarks/output/submission-bundle/guardbench-manifest.json +0 -414
  124. package/benchmarks/output/submission-bundle/guardbench-raw.json +0 -1279
  125. package/benchmarks/output/submission-bundle/guardbench-summary.json +0 -2123
  126. package/benchmarks/output/submission-bundle/schemas/guardbench-adapter-registry.schema.json +0 -69
  127. package/benchmarks/output/submission-bundle/schemas/guardbench-adapter-self-test.schema.json +0 -156
  128. package/benchmarks/output/submission-bundle/schemas/guardbench-conformance-card.schema.json +0 -184
  129. package/benchmarks/output/submission-bundle/schemas/guardbench-external-dry-run.schema.json +0 -74
  130. package/benchmarks/output/submission-bundle/schemas/guardbench-external-evidence.schema.json +0 -108
  131. package/benchmarks/output/submission-bundle/schemas/guardbench-external-run.schema.json +0 -160
  132. package/benchmarks/output/submission-bundle/schemas/guardbench-leaderboard.schema.json +0 -179
  133. package/benchmarks/output/submission-bundle/schemas/guardbench-manifest.schema.json +0 -213
  134. package/benchmarks/output/submission-bundle/schemas/guardbench-publication-verification.schema.json +0 -47
  135. package/benchmarks/output/submission-bundle/schemas/guardbench-raw.schema.json +0 -184
  136. package/benchmarks/output/submission-bundle/schemas/guardbench-submission-manifest.schema.json +0 -151
  137. package/benchmarks/output/submission-bundle/schemas/guardbench-summary.schema.json +0 -249
  138. package/benchmarks/output/submission-bundle/submission-manifest.json +0 -131
  139. package/benchmarks/output/submission-bundle/validation-report.json +0 -31
  140. package/benchmarks/output/summary.json +0 -2354
  141. package/benchmarks/perf-snapshot.js +0 -307
  142. package/benchmarks/perf.bench.js +0 -169
  143. package/benchmarks/public-paths.mjs +0 -97
  144. package/benchmarks/reference-results.js +0 -75
  145. package/benchmarks/report.js +0 -280
  146. package/benchmarks/run-external-guardbench.mjs +0 -303
  147. package/benchmarks/run.js +0 -735
  148. package/benchmarks/schemas/guardbench-adapter-registry.schema.json +0 -69
  149. package/benchmarks/schemas/guardbench-adapter-self-test.schema.json +0 -156
  150. package/benchmarks/schemas/guardbench-conformance-card.schema.json +0 -184
  151. package/benchmarks/schemas/guardbench-external-dry-run.schema.json +0 -74
  152. package/benchmarks/schemas/guardbench-external-evidence.schema.json +0 -108
  153. package/benchmarks/schemas/guardbench-external-run.schema.json +0 -160
  154. package/benchmarks/schemas/guardbench-leaderboard.schema.json +0 -179
  155. package/benchmarks/schemas/guardbench-manifest.schema.json +0 -213
  156. package/benchmarks/schemas/guardbench-publication-verification.schema.json +0 -47
  157. package/benchmarks/schemas/guardbench-raw.schema.json +0 -184
  158. package/benchmarks/schemas/guardbench-submission-manifest.schema.json +0 -151
  159. package/benchmarks/schemas/guardbench-summary.schema.json +0 -249
  160. package/benchmarks/snapshots/perf-0.22.2.json +0 -123
  161. package/benchmarks/snapshots/perf-0.23.0.json +0 -123
  162. package/benchmarks/validate-adapter-module.mjs +0 -107
  163. package/benchmarks/validate-adapter-registry.mjs +0 -145
  164. package/benchmarks/validate-adapter-self-test.mjs +0 -96
  165. package/benchmarks/validate-guardbench-artifacts.mjs +0 -400
  166. package/benchmarks/verify-external-evidence.mjs +0 -351
  167. package/benchmarks/verify-publication-artifacts.mjs +0 -309
  168. package/benchmarks/verify-submission-bundle.mjs +0 -172
  169. package/docs/AUDREY_PAPER_OUTLINE.md +0 -178
  170. package/docs/MEMORY_BENCHMARKING.md +0 -59
  171. package/docs/PRODUCTION_BACKLOG.md +0 -321
  172. package/docs/assets/audrey-feature-grid.jpg +0 -0
  173. package/docs/assets/audrey-logo.svg +0 -45
  174. package/docs/assets/audrey-wordmark.png +0 -0
  175. package/docs/paper/00-master.md +0 -48
  176. package/docs/paper/01-introduction.md +0 -27
  177. package/docs/paper/02-related-work.md +0 -47
  178. package/docs/paper/03-problem-definition.md +0 -108
  179. package/docs/paper/04-design.md +0 -164
  180. package/docs/paper/05-guardbench-spec.md +0 -412
  181. package/docs/paper/06-implementation.md +0 -115
  182. package/docs/paper/07-evaluation.md +0 -168
  183. package/docs/paper/08-discussion-limitations.md +0 -61
  184. package/docs/paper/09-conclusion.md +0 -11
  185. package/docs/paper/SUBMISSION_README.md +0 -162
  186. package/docs/paper/appendix-a-demo-transcript.md +0 -114
  187. package/docs/paper/arxiv-compile-report.schema.json +0 -116
  188. package/docs/paper/arxiv-source.schema.json +0 -61
  189. package/docs/paper/audrey-paper-v1.md +0 -1108
  190. package/docs/paper/browser-launch-plan.json +0 -209
  191. package/docs/paper/browser-launch-plan.schema.json +0 -100
  192. package/docs/paper/browser-launch-results.json +0 -86
  193. package/docs/paper/browser-launch-results.schema.json +0 -66
  194. package/docs/paper/claim-register.json +0 -138
  195. package/docs/paper/claim-register.schema.json +0 -81
  196. package/docs/paper/evidence-ledger.md +0 -103
  197. package/docs/paper/output/arxiv/README-arxiv.txt +0 -8
  198. package/docs/paper/output/arxiv/arxiv-manifest.json +0 -41
  199. package/docs/paper/output/arxiv/main.tex +0 -951
  200. package/docs/paper/output/arxiv/references.bib +0 -222
  201. package/docs/paper/output/arxiv-compile-report.json +0 -24
  202. package/docs/paper/output/submission-bundle/LICENSE +0 -21
  203. package/docs/paper/output/submission-bundle/README.md +0 -440
  204. package/docs/paper/output/submission-bundle/benchmarks/output/adapter-self-test/guardbench-adapter-self-test.json +0 -50
  205. package/docs/paper/output/submission-bundle/benchmarks/output/external/guardbench-external-dry-run.json +0 -69
  206. package/docs/paper/output/submission-bundle/benchmarks/output/external/guardbench-external-evidence.json +0 -56
  207. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-conformance-card.json +0 -63
  208. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-manifest.json +0 -414
  209. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-raw.json +0 -1279
  210. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-summary.json +0 -2123
  211. package/docs/paper/output/submission-bundle/benchmarks/output/leaderboard/guardbench-leaderboard.json +0 -93
  212. package/docs/paper/output/submission-bundle/benchmarks/output/leaderboard/guardbench-leaderboard.md +0 -7
  213. package/docs/paper/output/submission-bundle/benchmarks/output/submission-bundle/submission-manifest.json +0 -131
  214. package/docs/paper/output/submission-bundle/benchmarks/output/submission-bundle/validation-report.json +0 -31
  215. package/docs/paper/output/submission-bundle/benchmarks/output/summary.json +0 -2354
  216. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-adapter-registry.schema.json +0 -69
  217. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-adapter-self-test.schema.json +0 -156
  218. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-conformance-card.schema.json +0 -184
  219. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-external-dry-run.schema.json +0 -74
  220. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-external-evidence.schema.json +0 -108
  221. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-external-run.schema.json +0 -160
  222. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-leaderboard.schema.json +0 -179
  223. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-manifest.schema.json +0 -213
  224. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-publication-verification.schema.json +0 -47
  225. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-raw.schema.json +0 -184
  226. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-submission-manifest.schema.json +0 -151
  227. package/docs/paper/output/submission-bundle/benchmarks/schemas/guardbench-summary.schema.json +0 -249
  228. package/docs/paper/output/submission-bundle/docs/AUDREY_PAPER_OUTLINE.md +0 -178
  229. package/docs/paper/output/submission-bundle/docs/paper/00-master.md +0 -48
  230. package/docs/paper/output/submission-bundle/docs/paper/01-introduction.md +0 -27
  231. package/docs/paper/output/submission-bundle/docs/paper/02-related-work.md +0 -47
  232. package/docs/paper/output/submission-bundle/docs/paper/03-problem-definition.md +0 -108
  233. package/docs/paper/output/submission-bundle/docs/paper/04-design.md +0 -164
  234. package/docs/paper/output/submission-bundle/docs/paper/05-guardbench-spec.md +0 -412
  235. package/docs/paper/output/submission-bundle/docs/paper/06-implementation.md +0 -115
  236. package/docs/paper/output/submission-bundle/docs/paper/07-evaluation.md +0 -168
  237. package/docs/paper/output/submission-bundle/docs/paper/08-discussion-limitations.md +0 -61
  238. package/docs/paper/output/submission-bundle/docs/paper/09-conclusion.md +0 -11
  239. package/docs/paper/output/submission-bundle/docs/paper/SUBMISSION_README.md +0 -162
  240. package/docs/paper/output/submission-bundle/docs/paper/appendix-a-demo-transcript.md +0 -114
  241. package/docs/paper/output/submission-bundle/docs/paper/arxiv-compile-report.schema.json +0 -116
  242. package/docs/paper/output/submission-bundle/docs/paper/arxiv-source.schema.json +0 -61
  243. package/docs/paper/output/submission-bundle/docs/paper/audrey-paper-v1.md +0 -1108
  244. package/docs/paper/output/submission-bundle/docs/paper/browser-launch-plan.json +0 -209
  245. package/docs/paper/output/submission-bundle/docs/paper/browser-launch-plan.schema.json +0 -100
  246. package/docs/paper/output/submission-bundle/docs/paper/browser-launch-results.json +0 -86
  247. package/docs/paper/output/submission-bundle/docs/paper/browser-launch-results.schema.json +0 -66
  248. package/docs/paper/output/submission-bundle/docs/paper/claim-register.json +0 -138
  249. package/docs/paper/output/submission-bundle/docs/paper/claim-register.schema.json +0 -81
  250. package/docs/paper/output/submission-bundle/docs/paper/evidence-ledger.md +0 -103
  251. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/README-arxiv.txt +0 -8
  252. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/arxiv-manifest.json +0 -41
  253. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/main.tex +0 -951
  254. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/references.bib +0 -222
  255. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv-compile-report.json +0 -24
  256. package/docs/paper/output/submission-bundle/docs/paper/paper-submission-bundle.schema.json +0 -70
  257. package/docs/paper/output/submission-bundle/docs/paper/publication-pack.json +0 -81
  258. package/docs/paper/output/submission-bundle/docs/paper/publication-pack.schema.json +0 -60
  259. package/docs/paper/output/submission-bundle/docs/paper/references.bib +0 -222
  260. package/docs/paper/output/submission-bundle/package.json +0 -235
  261. package/docs/paper/output/submission-bundle/paper-submission-manifest.json +0 -379
  262. package/docs/paper/paper-submission-bundle.schema.json +0 -70
  263. package/docs/paper/publication-pack.json +0 -81
  264. package/docs/paper/publication-pack.schema.json +0 -60
  265. package/docs/paper/references.bib +0 -222
  266. package/scripts/audit-release-completion.mjs +0 -440
  267. package/scripts/create-arxiv-source.mjs +0 -368
  268. package/scripts/create-paper-submission-bundle.mjs +0 -214
  269. package/scripts/finalize-release.mjs +0 -601
  270. package/scripts/prepare-release-cut.mjs +0 -277
  271. package/scripts/publish-release-bundle.mjs +0 -252
  272. package/scripts/publish-release-github-api.mjs +0 -494
  273. package/scripts/run-vitest.mjs +0 -34
  274. package/scripts/smoke-cli.js +0 -92
  275. package/scripts/sync-paper-artifacts.mjs +0 -112
  276. package/scripts/verify-arxiv-compile.mjs +0 -476
  277. package/scripts/verify-arxiv-source.mjs +0 -224
  278. package/scripts/verify-browser-launch-plan.mjs +0 -254
  279. package/scripts/verify-browser-launch-results.mjs +0 -303
  280. package/scripts/verify-paper-artifacts.mjs +0 -798
  281. package/scripts/verify-paper-claims.mjs +0 -254
  282. package/scripts/verify-paper-submission-bundle.mjs +0 -218
  283. package/scripts/verify-publication-pack.mjs +0 -210
  284. package/scripts/verify-python-package.py +0 -201
  285. package/scripts/verify-release-readiness.mjs +0 -920
@@ -1,114 +0,0 @@
1
- # Appendix A. Repeated-Failure Demo Transcript
2
-
3
- This appendix records the qualitative demo used in Section 7. The demo creates an isolated temporary Audrey store, records one failed deploy command, stores the operational rule implied by the failure, and runs a pre-action guard check on the same command. It is the paper's artifact-grounded figure because it shows the central claim in one executable trace: memory changes the next tool action before the tool runs.
4
-
5
- ## Commands
6
-
7
- Build command:
8
-
9
- ```bash
10
- npm run build
11
- ```
12
-
13
- Run command:
14
-
15
- ```bash
16
- node dist/mcp-server/index.js demo --scenario repeated-failure
17
- ```
18
-
19
- ## Verbatim Transcript
20
-
21
- ```text
22
- Audrey Guard repeated-failure demo
23
-
24
- Memory store: [LOCAL-TEMP]/audrey-demo-AkCROa
25
- Step 1: the agent tries a deploy and hits a real setup failure.
26
- Step 2: Audrey stores the failure and the operational rule it implies.
27
- Lesson memory: 01KR491DG2YZHVEM79QVW5BHZA
28
-
29
- Step 3: a new preflight checks the same action before tool use.
30
-
31
- Audrey Guard: BLOCKED
32
-
33
- Reason: Blocked: this exact Bash action failed before. Stop: 3 memory reflexes, 2 blocking, 1 warning matched.
34
- Risk score: 0.90
35
-
36
- Evidence:
37
- - 01KR491DFZYZ20TFK71KJHC88F
38
- - 01KR491DG2YZHVEM79QVW5BHZA
39
- - failure:Bash:2026-05-08T17:09:22.047Z
40
-
41
- Recommended action:
42
- - Do not repeat the exact failed action until the prior error is understood or the command is changed.
43
- - Do not proceed until the high-severity memory warning is addressed.
44
- - Apply this must-follow rule before acting.
45
- - Mitigate this remembered risk before proceeding.
46
- - Before re-running Bash, check what changed since the last failure.
47
-
48
- Memory reflexes:
49
- - block: Apply this must-follow rule before acting. Before running npm run deploy, run npm run db:generate because Prisma client must be generated first.
50
- - block: Mitigate this remembered risk before proceeding. Before running npm run deploy, run npm run db:generate because Prisma client must be generated first.
51
- - warn: Before re-running Bash, check what changed since the last failure.
52
-
53
- Next: fix the warning and retry, or pass --override to allow this guard check.
54
-
55
- Impact:
56
- - 1 repeated failure prevented
57
- - 1 helpful memory validation recorded
58
- - 3 evidence ids attached
59
-
60
- Audrey saw the agent fail once.
61
- Audrey stopped it from failing twice.
62
- ```
63
-
64
- ## Line Annotations
65
-
66
- | Transcript Fragment | Demonstrates |
67
- |---|---|
68
- | `Audrey Guard repeated-failure demo` | The named scenario is the Guard demo, not a generic recall query. |
69
- | `Memory store: ...\audrey-demo-...` | The demo uses an isolated temporary memory store; IDs and temp path are run-specific. |
70
- | `Step 1: the agent tries a deploy...` | The first action is allowed to fail once, creating real operational evidence. |
71
- | `Step 2: Audrey stores the failure...` | The failed tool outcome is converted into memory state. |
72
- | `Lesson memory: ...` | The procedural lesson receives a concrete memory ID that can be cited later. |
73
- | `Step 3: a new preflight...` | The next decision occurs before tool use, which is the pre-action control boundary. |
74
- | `Audrey Guard: BLOCKED` | The controller returns an enforced block decision rather than retrieved context. |
75
- | `Reason: Blocked: this exact Bash action failed before...` | The decision combines exact repeated-failure matching with reflex summary counts. |
76
- | `Risk score: 0.90` | The guard exposes a numeric risk score in the decision object. |
77
- | `Evidence:` and the three evidence rows | The block is auditable: prior event, lesson memory, and failure-class evidence are attached. |
78
- | `Recommended action:` rows | The guard returns concrete next actions rather than only a warning sentence. |
79
- | `Memory reflexes:` rows | Preflight warnings are converted into block and warn reflexes with operational wording. |
80
- | `Next: fix the warning...` | The CLI output preserves an explicit override path while defaulting to prevention. |
81
- | `Impact:` rows | The loop records prevention, validation, and attached-evidence accounting. |
82
- | `Audrey saw the agent fail once.` | The system observes failure instead of pretending it can prevent first-time unknown errors. |
83
- | `Audrey stopped it from failing twice.` | The central behavior: memory changes the next tool action. |
84
-
85
- ## How to Reproduce
86
-
87
- Prerequisites:
88
-
89
- - Node.js 20 or newer.
90
- - Audrey dependencies installed with `npm install`.
91
- - A clean or dirty checkout is acceptable; the demo uses a temporary mock-provider store and does not write to the user's normal Audrey data directory.
92
-
93
- Commands:
94
-
95
- ```bash
96
- npm run build
97
- node dist/mcp-server/index.js demo --scenario repeated-failure
98
- ```
99
-
100
- Expected output shape:
101
-
102
- - The command prints `Audrey Guard repeated-failure demo`.
103
- - It reports a temporary `Memory store` path.
104
- - It prints a run-specific `Lesson memory` ID.
105
- - It prints `Audrey Guard: BLOCKED`.
106
- - It prints `Risk score: 0.90`.
107
- - It lists at least one prior failure evidence ID, one lesson memory ID, and one `failure:Bash:<timestamp>` evidence ID.
108
- - It reports two blocking reflexes, one warning reflex, one repeated failure prevented, one helpful memory validation recorded, and three evidence IDs attached.
109
-
110
- Run-specific values:
111
-
112
- - The temp directory suffix changes on each run.
113
- - Memory IDs change on each run.
114
- - The timestamp inside the `failure:Bash:<timestamp>` evidence ID changes on each run.
@@ -1,116 +0,0 @@
1
- {
2
- "$schema": "https://json-schema.org/draft/2020-12/schema",
3
- "$id": "https://audrey-memory.org/schemas/audrey-arxiv-compile-report.schema.json",
4
- "title": "Audrey arXiv Compile Report",
5
- "type": "object",
6
- "additionalProperties": false,
7
- "required": [
8
- "schemaVersion",
9
- "suite",
10
- "generatedAt",
11
- "source",
12
- "outputDir",
13
- "status",
14
- "compiler",
15
- "outputPdf",
16
- "outputPdfSha256",
17
- "logFile",
18
- "blockers",
19
- "failures"
20
- ],
21
- "properties": {
22
- "schemaVersion": { "const": "1.0.0" },
23
- "suite": { "const": "Audrey arXiv compile check" },
24
- "generatedAt": { "type": "string", "minLength": 1 },
25
- "source": {
26
- "anyOf": [
27
- { "type": "null" },
28
- {
29
- "type": "object",
30
- "additionalProperties": false,
31
- "required": [
32
- "sourceDir",
33
- "manifest",
34
- "manifestSha256",
35
- "mainTex",
36
- "mainTexSha256",
37
- "referencesBib",
38
- "referencesBibSha256"
39
- ],
40
- "properties": {
41
- "sourceDir": { "type": "string", "minLength": 1 },
42
- "manifest": { "type": "string", "minLength": 1 },
43
- "manifestSha256": { "$ref": "#/$defs/sha256" },
44
- "mainTex": { "type": "string", "minLength": 1 },
45
- "mainTexSha256": { "$ref": "#/$defs/sha256" },
46
- "referencesBib": { "type": "string", "minLength": 1 },
47
- "referencesBibSha256": { "$ref": "#/$defs/sha256" }
48
- }
49
- }
50
- ]
51
- },
52
- "outputDir": { "type": "string", "minLength": 1 },
53
- "status": { "enum": ["passed", "failed", "toolchain-missing"] },
54
- "compiler": {
55
- "anyOf": [
56
- { "type": "null" },
57
- {
58
- "type": "object",
59
- "additionalProperties": false,
60
- "required": ["name", "stages"],
61
- "properties": {
62
- "name": { "enum": ["tectonic", "latexmk", "pdflatex+bibtex", "uvx-tecto"] },
63
- "stages": {
64
- "type": "array",
65
- "minItems": 1,
66
- "items": {
67
- "type": "object",
68
- "additionalProperties": false,
69
- "required": ["command", "args"],
70
- "properties": {
71
- "command": { "type": "string", "minLength": 1 },
72
- "args": {
73
- "type": "array",
74
- "items": { "type": "string" }
75
- }
76
- }
77
- }
78
- }
79
- }
80
- }
81
- ]
82
- },
83
- "outputPdf": {
84
- "anyOf": [
85
- { "type": "null" },
86
- { "type": "string", "minLength": 1 }
87
- ]
88
- },
89
- "outputPdfSha256": {
90
- "anyOf": [
91
- { "type": "null" },
92
- { "$ref": "#/$defs/sha256" }
93
- ]
94
- },
95
- "logFile": {
96
- "anyOf": [
97
- { "type": "null" },
98
- { "type": "string", "minLength": 1 }
99
- ]
100
- },
101
- "blockers": {
102
- "type": "array",
103
- "items": { "type": "string" }
104
- },
105
- "failures": {
106
- "type": "array",
107
- "items": { "type": "string" }
108
- }
109
- },
110
- "$defs": {
111
- "sha256": {
112
- "type": "string",
113
- "pattern": "^[a-f0-9]{64}$"
114
- }
115
- }
116
- }
@@ -1,61 +0,0 @@
1
- {
2
- "$schema": "https://json-schema.org/draft/2020-12/schema",
3
- "$id": "https://audrey-memory.org/schemas/audrey-arxiv-source-manifest.schema.json",
4
- "title": "Audrey arXiv Source Manifest",
5
- "type": "object",
6
- "additionalProperties": false,
7
- "required": ["schemaVersion", "suite", "generatedAt", "sourceMarkdown", "publicationPack", "sourceHashes", "files", "tex"],
8
- "properties": {
9
- "schemaVersion": { "const": "1.0.0" },
10
- "suite": { "const": "Audrey arXiv source package" },
11
- "generatedAt": { "type": "string", "minLength": 1 },
12
- "sourceMarkdown": { "type": "string", "minLength": 1 },
13
- "publicationPack": { "type": "string", "minLength": 1 },
14
- "sourceHashes": {
15
- "type": "object",
16
- "additionalProperties": false,
17
- "required": ["sourceMarkdown", "publicationPack", "referencesBib"],
18
- "properties": {
19
- "sourceMarkdown": { "$ref": "#/$defs/sha256" },
20
- "publicationPack": { "$ref": "#/$defs/sha256" },
21
- "referencesBib": { "$ref": "#/$defs/sha256" }
22
- }
23
- },
24
- "files": {
25
- "type": "array",
26
- "minItems": 1,
27
- "items": { "$ref": "#/$defs/fileRecord" }
28
- },
29
- "tex": {
30
- "type": "object",
31
- "additionalProperties": false,
32
- "required": ["mainFile", "titleEntry", "abstractEntry", "citationCount", "bibEntryCount", "codeBlocks", "tableBlocks"],
33
- "properties": {
34
- "mainFile": { "const": "main.tex" },
35
- "titleEntry": { "const": "arxiv-title" },
36
- "abstractEntry": { "const": "arxiv-abstract" },
37
- "citationCount": { "type": "integer", "minimum": 1 },
38
- "bibEntryCount": { "type": "integer", "minimum": 1 },
39
- "codeBlocks": { "type": "integer", "minimum": 0 },
40
- "tableBlocks": { "type": "integer", "minimum": 0 }
41
- }
42
- }
43
- },
44
- "$defs": {
45
- "fileRecord": {
46
- "type": "object",
47
- "additionalProperties": false,
48
- "required": ["path", "source", "bytes", "sha256"],
49
- "properties": {
50
- "path": { "type": "string", "minLength": 1 },
51
- "source": { "type": "string", "minLength": 1 },
52
- "bytes": { "type": "integer", "minimum": 0 },
53
- "sha256": { "$ref": "#/$defs/sha256" }
54
- }
55
- },
56
- "sha256": {
57
- "type": "string",
58
- "pattern": "^[a-f0-9]{64}$"
59
- }
60
- }
61
- }