@cursor/july 0.1.19 → 0.1.21

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 (217) hide show
  1. package/AGENTS.md +10 -0
  2. package/dist/bin/agent-serve.js +0 -0
  3. package/dist/channels/github/instrument.d.ts +20 -0
  4. package/dist/channels/github/instrument.d.ts.map +1 -0
  5. package/dist/channels/slack/api.d.ts +2 -0
  6. package/dist/channels/slack/api.d.ts.map +1 -1
  7. package/dist/channels/slack/api.js +3 -0
  8. package/dist/channels/slack/bot-mentions.d.ts +106 -0
  9. package/dist/channels/slack/bot-mentions.d.ts.map +1 -0
  10. package/dist/channels/slack/bot-mentions.js +243 -0
  11. package/dist/channels/slack/dispatch.d.ts +6 -0
  12. package/dist/channels/slack/dispatch.d.ts.map +1 -1
  13. package/dist/channels/slack/dispatch.js +24 -3
  14. package/dist/channels/slack/inbound.d.ts +10 -1
  15. package/dist/channels/slack/inbound.d.ts.map +1 -1
  16. package/dist/channels/slack/inbound.js +12 -4
  17. package/dist/channels/slack/index.d.ts +1 -0
  18. package/dist/channels/slack/index.d.ts.map +1 -1
  19. package/dist/channels/slack/index.js +1 -0
  20. package/dist/channels/slack/slack-channel.d.ts.map +1 -1
  21. package/dist/channels/slack/slack-channel.js +61 -22
  22. package/dist/channels/slack/types.d.ts +58 -0
  23. package/dist/channels/slack/types.d.ts.map +1 -1
  24. package/dist/docs/404.html +1 -1
  25. package/dist/docs/ab.html +2 -2
  26. package/dist/docs/assets/{app.CrsWMchO.js → app.jDxLzWv4.js} +1 -1
  27. package/dist/docs/assets/chunks/@localSearchIndexroot.DoJHJjqF.js +1 -0
  28. package/dist/docs/assets/chunks/{VPLocalSearchBox.D1JqzSh8.js → VPLocalSearchBox.Y6bDR1-a.js} +1 -1
  29. package/dist/docs/assets/chunks/{theme.DaBvZYwl.js → theme.CLazCWlJ.js} +2 -2
  30. package/dist/docs/building-with-agents.html +2 -2
  31. package/dist/docs/concepts.html +2 -2
  32. package/dist/docs/deployment.html +2 -2
  33. package/dist/docs/evals.html +2 -2
  34. package/dist/docs/example-agents/approval-buddy.html +2 -2
  35. package/dist/docs/example-agents/benny.html +2 -2
  36. package/dist/docs/example-agents/bugbot.html +2 -2
  37. package/dist/docs/example-agents/codebase-wiki.html +2 -2
  38. package/dist/docs/example-agents/codeowners-review.html +2 -2
  39. package/dist/docs/example-agents/concierge.html +2 -2
  40. package/dist/docs/example-agents/fsd.html +2 -2
  41. package/dist/docs/example-agents/index.html +2 -2
  42. package/dist/docs/example-agents/knowledge-base.html +2 -2
  43. package/dist/docs/example-agents/oncall.html +2 -2
  44. package/dist/docs/example-agents/security-reviewer.html +2 -2
  45. package/dist/docs/example-agents/slack-agent.html +2 -2
  46. package/dist/docs/example-agents/weather-agent.html +2 -2
  47. package/dist/docs/guides/agent-to-agent.html +2 -2
  48. package/dist/docs/guides/cloud-runtime.html +2 -2
  49. package/dist/docs/guides/github.html +2 -2
  50. package/dist/docs/guides/human-in-the-loop.html +2 -2
  51. package/dist/docs/guides/mcp-oauth.html +2 -2
  52. package/dist/docs/guides/slack.html +2 -2
  53. package/dist/docs/guides/webhooks.html +2 -2
  54. package/dist/docs/hillclimbing.html +2 -2
  55. package/dist/docs/index.html +2 -2
  56. package/dist/docs/quickstart.html +2 -2
  57. package/dist/docs/reference/agent-config.html +2 -2
  58. package/dist/docs/reference/channels.html +2 -2
  59. package/dist/docs/reference/cli.html +2 -2
  60. package/dist/docs/reference/connections.html +2 -2
  61. package/dist/docs/reference/hooks.html +2 -2
  62. package/dist/docs/reference/http-api.html +2 -2
  63. package/dist/docs/reference/instructions.html +2 -2
  64. package/dist/docs/reference/playground.html +2 -2
  65. package/dist/docs/reference/project-layout.html +2 -2
  66. package/dist/docs/reference/prompt.html +2 -2
  67. package/dist/docs/reference/schedules.html +2 -2
  68. package/dist/docs/reference/sessions.html +2 -2
  69. package/dist/docs/reference/skills.html +2 -2
  70. package/dist/docs/reference/subagents.html +2 -2
  71. package/dist/docs/reference/tools.html +2 -2
  72. package/dist/docs/scaffolding-agents.html +2 -2
  73. package/dist/docs/storage.html +2 -2
  74. package/dist/docs/troubleshooting.html +2 -2
  75. package/dist/internal/json-dir-store.d.ts +32 -0
  76. package/dist/internal/json-dir-store.d.ts.map +1 -0
  77. package/dist/internal/json-dir-store.js +100 -0
  78. package/dist/internal/resolved-connections.d.ts +1 -1
  79. package/dist/internal/resolved-connections.d.ts.map +1 -1
  80. package/dist/internal/resolved-connections.js +10 -0
  81. package/dist/internal/session-engine.d.ts +4 -1
  82. package/dist/internal/session-engine.d.ts.map +1 -1
  83. package/dist/internal/session-engine.js +13 -2
  84. package/dist/playground/assets/index-CjOQ4hN9.css +1 -0
  85. package/dist/playground/assets/{index-C2SU2xV5.js → index-dshZQJCp.js} +46 -46
  86. package/dist/playground/index.html +2 -2
  87. package/dist/types.d.ts +7 -0
  88. package/dist/types.d.ts.map +1 -1
  89. package/dist/types.js +9 -0
  90. package/package.json +25 -25
  91. package/skills/create-agent/SKILL.md +36 -4
  92. package/skills/evals/SKILL.md +3 -0
  93. package/skills/framework-map/SKILL.md +14 -0
  94. package/skills/github/SKILL.md +6 -1
  95. package/skills/hillclimb/SKILL.md +28 -7
  96. package/src/bin/agent-serve.version.test.ts +62 -0
  97. package/src/channels/github/api.test.ts +64 -0
  98. package/src/channels/github/auth.test.ts +105 -0
  99. package/src/channels/github/cursor-account.test.ts +204 -0
  100. package/src/channels/github/forward.test.ts +457 -0
  101. package/src/channels/github/github.test.ts +937 -0
  102. package/src/channels/github/replay.test.ts +179 -0
  103. package/src/channels/slack/api.post-message.test.ts +148 -0
  104. package/src/channels/slack/api.ts +5 -0
  105. package/src/channels/slack/approvals.test.ts +328 -0
  106. package/src/channels/slack/block-actions.test.ts +452 -0
  107. package/src/channels/slack/bot-mentions.test.ts +217 -0
  108. package/src/channels/slack/bot-mentions.ts +315 -0
  109. package/src/channels/slack/channel-watch.test.ts +363 -0
  110. package/src/channels/slack/cursor-account.test.ts +253 -0
  111. package/src/channels/slack/defaults.final-post.test.ts +182 -0
  112. package/src/channels/slack/dispatch.test.ts +795 -0
  113. package/src/channels/slack/dispatch.ts +27 -1
  114. package/src/channels/slack/eval-directive.test.ts +273 -0
  115. package/src/channels/slack/inbound.ts +25 -4
  116. package/src/channels/slack/index.ts +1 -0
  117. package/src/channels/slack/message-body.test.ts +54 -0
  118. package/src/channels/slack/nudge-store.test.ts +143 -0
  119. package/src/channels/slack/slack-channel.ts +66 -8
  120. package/src/channels/slack/slack.test.ts +391 -0
  121. package/src/channels/slack/stop.test.ts +23 -0
  122. package/src/channels/slack/thread-context.test.ts +202 -0
  123. package/src/channels/slack/types.ts +60 -0
  124. package/src/evals/assertions.test.ts +580 -0
  125. package/src/evals/expect.test.ts +144 -0
  126. package/src/evals/judge.test.ts +181 -0
  127. package/src/evals/loaders.test.ts +132 -0
  128. package/src/evals/matchers.test.ts +95 -0
  129. package/src/evals/reporters.test.ts +303 -0
  130. package/src/evals/run-facts.test.ts +259 -0
  131. package/src/internal/ab-snapshot.test.ts +325 -0
  132. package/src/internal/approval-gate.test.ts +49 -0
  133. package/src/internal/approvals.integration.test.ts +383 -0
  134. package/src/internal/authored-loaders.test.ts +31 -0
  135. package/src/internal/builtin-tools/reminders.test.ts +201 -0
  136. package/src/internal/channel-route-schema.test.ts +294 -0
  137. package/src/internal/chat-attach.test.ts +262 -0
  138. package/src/internal/cli-deploy.test.ts +1991 -0
  139. package/src/internal/cli-mcp.test.ts +789 -0
  140. package/src/internal/cli-skills.test.ts +133 -0
  141. package/src/internal/cli-slack.test.ts +1647 -0
  142. package/src/internal/cloud-merge.test.ts +74 -0
  143. package/src/internal/cron.test.ts +22 -0
  144. package/src/internal/cursor/account-mcp.test.ts +807 -0
  145. package/src/internal/cursor/backend-client.test.ts +591 -0
  146. package/src/internal/cursor/credentials.test.ts +351 -0
  147. package/src/internal/cursor/github-credentials.test.ts +136 -0
  148. package/src/internal/cursor-account-mcp-auth.test.ts +310 -0
  149. package/src/internal/cursor-account.integration.test.ts +441 -0
  150. package/src/internal/cursor-event-relay.test.ts +746 -0
  151. package/src/internal/cursor-github-credentials.integration.test.ts +271 -0
  152. package/src/internal/cursor-slack-relay.test.ts +525 -0
  153. package/src/internal/deploy-source.test.ts +111 -0
  154. package/src/internal/discovery.builtin-tools.test.ts +94 -0
  155. package/src/internal/discovery.concurrency.test.ts +60 -0
  156. package/src/internal/discovery.cursor-account.test.ts +133 -0
  157. package/src/internal/discovery.cwd.test.ts +83 -0
  158. package/src/internal/discovery.hosting.test.ts +80 -0
  159. package/src/internal/discovery.identity.test.ts +44 -0
  160. package/src/internal/docs-site.test.ts +66 -0
  161. package/src/internal/duration.test.ts +29 -0
  162. package/src/internal/eval-judge-model.test.ts +187 -0
  163. package/src/internal/eval-run-store.cancel.test.ts +142 -0
  164. package/src/internal/eval-run-store.storage.test.ts +211 -0
  165. package/src/internal/eval-runner.http.test.ts +403 -0
  166. package/src/internal/eval-runner.run.test.ts +928 -0
  167. package/src/internal/evals-client.test.ts +307 -0
  168. package/src/internal/event-mapper.test.ts +243 -0
  169. package/src/internal/github-fanout.test.ts +213 -0
  170. package/src/internal/handleAgentServeTrigger.test.ts +179 -0
  171. package/src/internal/host-kv.test.ts +82 -0
  172. package/src/internal/host-platforms.test.ts +126 -0
  173. package/src/internal/http-channel.test.ts +402 -0
  174. package/src/internal/init-project.test.ts +269 -0
  175. package/src/internal/install-cursor-skills.test.ts +262 -0
  176. package/src/internal/local-env.test.ts +100 -0
  177. package/src/internal/log-ring.test.ts +31 -0
  178. package/src/internal/logs-client.test.ts +350 -0
  179. package/src/internal/mcp-endpoint.test.ts +436 -0
  180. package/src/internal/mcp-host.test.ts +298 -0
  181. package/src/internal/mcp-oauth.test.ts +148 -0
  182. package/src/internal/net.test.ts +17 -0
  183. package/src/internal/peer-connections.test.ts +128 -0
  184. package/src/internal/peer-mcp.integration.test.ts +289 -0
  185. package/src/internal/playground/toolchain.test.ts +53 -0
  186. package/src/internal/playground-cli.test.ts +187 -0
  187. package/src/internal/playground-proxy.test.ts +376 -0
  188. package/src/internal/prompt-context.integration.test.ts +232 -0
  189. package/src/internal/prompt-context.test.ts +127 -0
  190. package/src/internal/reminder-runner.test.ts +390 -0
  191. package/src/internal/reminder-store.test.ts +53 -0
  192. package/src/internal/request-headers.test.ts +27 -0
  193. package/src/internal/resolve-prod-target.test.ts +787 -0
  194. package/src/internal/resolved-connections.test.ts +295 -0
  195. package/src/internal/resolved-connections.ts +18 -1
  196. package/src/internal/router.test.ts +57 -0
  197. package/src/internal/sdk-runner.test.ts +290 -0
  198. package/src/internal/session-engine.coalesce.test.ts +169 -0
  199. package/src/internal/session-engine.concurrency.test.ts +250 -0
  200. package/src/internal/session-engine.host-oauth-mcp.test.ts +110 -0
  201. package/src/internal/session-engine.interrupt.test.ts +577 -0
  202. package/src/internal/session-engine.storage.test.ts +547 -0
  203. package/src/internal/session-engine.ts +15 -1
  204. package/src/internal/session-urls.test.ts +28 -0
  205. package/src/internal/sessions-client.test.ts +518 -0
  206. package/src/internal/storage-coordinator.test.ts +517 -0
  207. package/src/internal/tool-call.test.ts +458 -0
  208. package/src/internal/tool-result.test.ts +52 -0
  209. package/src/internal/trajectory.approvals.test.ts +83 -0
  210. package/src/internal/trajectory.subagents.test.ts +198 -0
  211. package/src/internal/turn-governor.test.ts +137 -0
  212. package/src/internal/update-check.test.ts +485 -0
  213. package/src/internal/workspace.test.ts +81 -0
  214. package/src/storage-backends/cursor-hosted.test.ts +121 -0
  215. package/src/types.ts +12 -0
  216. package/dist/docs/assets/chunks/@localSearchIndexroot.BnHRjfoe.js +0 -1
  217. package/dist/playground/assets/index-CidizGZv.css +0 -1
@@ -0,0 +1,325 @@
1
+ import { describe, expect, it } from "vitest";
2
+ import { defineAB, emptyABMetrics, splitAlways, splitNone } from "../ab.js";
3
+ import type { SessionEvent, SessionRecord, SessionSummary } from "../types.js";
4
+ import { foldABStream } from "./ab-fold.js";
5
+ import { buildABSnapshot } from "./ab-snapshot.js";
6
+
7
+ function event(
8
+ type: SessionEvent["type"],
9
+ index: number,
10
+ at: string,
11
+ data: Record<string, unknown> = {},
12
+ sessionId = "ses_1"
13
+ ): SessionEvent {
14
+ return {
15
+ type,
16
+ index,
17
+ at,
18
+ sessionId,
19
+ data,
20
+ } as SessionEvent;
21
+ }
22
+
23
+ describe("foldABStream", () => {
24
+ it("assigns arms and accumulates built-in metrics", () => {
25
+ const folded = foldABStream([
26
+ event("session.started", 0, "2026-01-01T00:00:00.000Z", {
27
+ channelId: "http",
28
+ }),
29
+ event("ab.assigned", 1, "2026-01-01T00:00:00.100Z", {
30
+ experiment: "exp",
31
+ variant: "treatment",
32
+ }),
33
+ event("turn.started", 2, "2026-01-01T00:00:01.000Z"),
34
+ event("action.result", 3, "2026-01-01T00:00:01.500Z", {
35
+ callId: "c1",
36
+ toolName: "get_weather",
37
+ isError: false,
38
+ }),
39
+ event("turn.completed", 4, "2026-01-01T00:00:02.000Z", {
40
+ usage: {
41
+ inputTokens: 10,
42
+ outputTokens: 5,
43
+ cacheReadTokens: 0,
44
+ cacheWriteTokens: 0,
45
+ },
46
+ }),
47
+ ]);
48
+ expect(folded.arms.get("exp")).toBe("treatment");
49
+ expect(folded.metrics.get("exp")?.metrics).toMatchObject({
50
+ turns: 1,
51
+ toolCalls: 1,
52
+ inputTokens: 10,
53
+ outputTokens: 5,
54
+ wallTimeMs: 1000,
55
+ });
56
+ });
57
+
58
+ it("sticky skip leaves no metrics bucket", () => {
59
+ const folded = foldABStream([
60
+ event("ab.assigned", 0, "2026-01-01T00:00:00.000Z", {
61
+ experiment: "exp",
62
+ variant: null,
63
+ }),
64
+ event("turn.completed", 1, "2026-01-01T00:00:01.000Z", {}),
65
+ ]);
66
+ expect(folded.arms.get("exp")).toBeNull();
67
+ expect(folded.metrics.has("exp")).toBe(false);
68
+ });
69
+ });
70
+
71
+ describe("buildABSnapshot", () => {
72
+ it("omits purpose:eval sessions even if the log has ab.assigned", async () => {
73
+ const def = defineAB({
74
+ name: "exp",
75
+ variants: { control: {}, treatment: {} },
76
+ split: splitAlways("treatment"),
77
+ onSample: () => {},
78
+ });
79
+ const snap = await buildABSnapshot({
80
+ experiments: [{ name: "exp", definition: def }],
81
+ agentName: "weather",
82
+ sessions: [
83
+ {
84
+ sessionId: "ses_eval",
85
+ channelId: "http",
86
+ mode: "chat",
87
+ turnCount: 1,
88
+ running: false,
89
+ createdAt: "2026-01-01T00:00:00.000Z",
90
+ updatedAt: "2026-01-01T00:00:01.000Z",
91
+ continuationToken: null,
92
+ },
93
+ ],
94
+ loadRecord: async () =>
95
+ ({
96
+ sessionId: "ses_eval",
97
+ channelId: "http",
98
+ mode: "chat",
99
+ purpose: "eval",
100
+ continuationKey: "k",
101
+ auth: null,
102
+ turnCount: 1,
103
+ createdAt: "2026-01-01T00:00:00.000Z",
104
+ updatedAt: "2026-01-01T00:00:01.000Z",
105
+ }) satisfies SessionRecord,
106
+ loadEvents: async () => [
107
+ event(
108
+ "ab.assigned",
109
+ 0,
110
+ "2026-01-01T00:00:00.100Z",
111
+ { experiment: "exp", variant: "treatment" },
112
+ "ses_eval"
113
+ ),
114
+ event(
115
+ "turn.completed",
116
+ 1,
117
+ "2026-01-01T00:00:01.000Z",
118
+ {
119
+ usage: {
120
+ inputTokens: 1,
121
+ outputTokens: 1,
122
+ cacheReadTokens: 0,
123
+ cacheWriteTokens: 0,
124
+ },
125
+ },
126
+ "ses_eval"
127
+ ),
128
+ ],
129
+ sessionInfo: (record) => ({
130
+ id: record.sessionId,
131
+ channelId: record.channelId,
132
+ mode: record.mode,
133
+ purpose: record.purpose,
134
+ auth: record.auth,
135
+ continuationKey: record.continuationKey,
136
+ }),
137
+ logger: () => {},
138
+ });
139
+ expect(snap.sessions).toEqual([]);
140
+ expect(snap.experiments[0]!.arms.every((a) => a.sessions === 0)).toBe(true);
141
+ });
142
+
143
+ it("aggregates per arm and lists per-session metrics", async () => {
144
+ const def = defineAB({
145
+ name: "exp",
146
+ variants: {
147
+ control: { label: "Baseline" },
148
+ treatment: { label: "Shorter" },
149
+ },
150
+ split: splitAlways("treatment"),
151
+ derive: {
152
+ weather: (e) =>
153
+ e.type === "action.result" && e.data.toolName === "get_weather"
154
+ ? 1
155
+ : null,
156
+ },
157
+ onSample: () => {},
158
+ });
159
+ const skipDef = defineAB({
160
+ name: "paused",
161
+ variants: { a: {}, b: {} },
162
+ split: splitNone(),
163
+ onSample: () => {},
164
+ });
165
+
166
+ const eventsBySession: Record<string, SessionEvent[]> = {
167
+ ses_a: [
168
+ event(
169
+ "ab.assigned",
170
+ 0,
171
+ "2026-01-01T00:00:00.000Z",
172
+ { experiment: "exp", variant: "treatment" },
173
+ "ses_a"
174
+ ),
175
+ event(
176
+ "ab.assigned",
177
+ 1,
178
+ "2026-01-01T00:00:00.050Z",
179
+ { experiment: "paused", variant: null },
180
+ "ses_a"
181
+ ),
182
+ event("turn.started", 2, "2026-01-01T00:00:01.000Z", {}, "ses_a"),
183
+ event(
184
+ "action.result",
185
+ 3,
186
+ "2026-01-01T00:00:01.200Z",
187
+ { callId: "c1", toolName: "get_weather", isError: false },
188
+ "ses_a"
189
+ ),
190
+ event(
191
+ "turn.completed",
192
+ 4,
193
+ "2026-01-01T00:00:02.000Z",
194
+ {
195
+ usage: {
196
+ inputTokens: 4,
197
+ outputTokens: 2,
198
+ cacheReadTokens: 0,
199
+ cacheWriteTokens: 0,
200
+ },
201
+ },
202
+ "ses_a"
203
+ ),
204
+ ],
205
+ ses_b: [
206
+ event(
207
+ "ab.assigned",
208
+ 0,
209
+ "2026-01-01T00:00:00.000Z",
210
+ { experiment: "exp", variant: "control" },
211
+ "ses_b"
212
+ ),
213
+ event("turn.started", 1, "2026-01-01T00:00:01.000Z", {}, "ses_b"),
214
+ event(
215
+ "turn.completed",
216
+ 2,
217
+ "2026-01-01T00:00:01.500Z",
218
+ {
219
+ usage: {
220
+ inputTokens: 1,
221
+ outputTokens: 1,
222
+ cacheReadTokens: 0,
223
+ cacheWriteTokens: 0,
224
+ },
225
+ },
226
+ "ses_b"
227
+ ),
228
+ ],
229
+ };
230
+
231
+ const summaries: SessionSummary[] = [
232
+ {
233
+ sessionId: "ses_a",
234
+ channelId: "http",
235
+ mode: "chat",
236
+ turnCount: 1,
237
+ running: false,
238
+ createdAt: "2026-01-01T00:00:00.000Z",
239
+ updatedAt: "2026-01-01T00:00:02.000Z",
240
+ continuationToken: null,
241
+ },
242
+ {
243
+ sessionId: "ses_b",
244
+ channelId: "http",
245
+ mode: "chat",
246
+ turnCount: 1,
247
+ running: false,
248
+ createdAt: "2026-01-01T00:00:00.000Z",
249
+ updatedAt: "2026-01-01T00:00:01.500Z",
250
+ continuationToken: null,
251
+ },
252
+ ];
253
+
254
+ const snap = await buildABSnapshot({
255
+ experiments: [
256
+ { name: "exp", definition: def },
257
+ { name: "paused", definition: skipDef },
258
+ ],
259
+ agentName: "weather",
260
+ sessions: summaries,
261
+ loadRecord: async (sessionId) =>
262
+ ({
263
+ sessionId,
264
+ channelId: "http",
265
+ mode: "chat",
266
+ continuationKey: "k",
267
+ auth: null,
268
+ turnCount: 1,
269
+ createdAt: "2026-01-01T00:00:00.000Z",
270
+ updatedAt: "2026-01-01T00:00:02.000Z",
271
+ }) satisfies SessionRecord,
272
+ loadEvents: async (sessionId) => eventsBySession[sessionId] ?? [],
273
+ sessionInfo: (record) => ({
274
+ id: record.sessionId,
275
+ channelId: record.channelId,
276
+ mode: record.mode,
277
+ auth: record.auth,
278
+ continuationKey: record.continuationKey,
279
+ }),
280
+ logger: () => {},
281
+ });
282
+
283
+ expect(snap.sessions).toHaveLength(2);
284
+ const sesA = snap.sessions[0]!;
285
+ expect(sesA.assignments).toEqual(
286
+ expect.arrayContaining([
287
+ expect.objectContaining({
288
+ experiment: "exp",
289
+ variant: "treatment",
290
+ variantLabel: "Shorter",
291
+ metrics: expect.objectContaining({
292
+ turns: 1,
293
+ toolCalls: 1,
294
+ custom: { weather: 1 },
295
+ }),
296
+ }),
297
+ expect.objectContaining({
298
+ experiment: "paused",
299
+ variant: null,
300
+ }),
301
+ ])
302
+ );
303
+
304
+ const exp = snap.experiments.find((e) => e.name === "exp")!;
305
+ expect(exp.skipped).toBe(0);
306
+ const treatment = exp.arms.find((a) => a.variant === "treatment")!;
307
+ const control = exp.arms.find((a) => a.variant === "control")!;
308
+ expect(treatment.sessions).toBe(1);
309
+ expect(treatment.metrics.toolCalls).toBe(1);
310
+ expect(treatment.metrics.custom.weather).toBe(1);
311
+ expect(control.sessions).toBe(1);
312
+ expect(control.metrics.turns).toBe(1);
313
+ expect(control.metrics).toMatchObject({
314
+ ...emptyABMetrics(),
315
+ turns: 1,
316
+ inputTokens: 1,
317
+ outputTokens: 1,
318
+ wallTimeMs: 500,
319
+ });
320
+
321
+ const paused = snap.experiments.find((e) => e.name === "paused")!;
322
+ expect(paused.skipped).toBe(1);
323
+ expect(paused.arms.every((a) => a.sessions === 0)).toBe(true);
324
+ });
325
+ });
@@ -0,0 +1,49 @@
1
+ import { describe, expect, it } from "vitest";
2
+ import { ApprovalGate } from "./approval-gate.js";
3
+
4
+ describe("ApprovalGate", () => {
5
+ it("parks until resolve(approve)", async () => {
6
+ const gate = new ApprovalGate();
7
+ const waiting = gate.wait({
8
+ sessionId: "s1",
9
+ callId: "c1",
10
+ turnId: "turn_1",
11
+ toolName: "post_weather_alert",
12
+ args: { city: "NYC" },
13
+ requestedAt: new Date().toISOString(),
14
+ });
15
+ expect(gate.has("s1", "c1")).toBe(true);
16
+ expect(gate.list("s1")).toHaveLength(1);
17
+ expect(gate.resolve("s1", "c1", "approve")).toBe(true);
18
+ await expect(waiting).resolves.toEqual({ decision: "approve" });
19
+ expect(gate.has("s1", "c1")).toBe(false);
20
+ });
21
+
22
+ it("returns false when no live waiter", () => {
23
+ const gate = new ApprovalGate();
24
+ expect(gate.resolve("missing", "c1", "deny")).toBe(false);
25
+ });
26
+
27
+ it("cancelSession denies all waiters for a session", async () => {
28
+ const gate = new ApprovalGate();
29
+ const first = gate.wait({
30
+ sessionId: "s1",
31
+ callId: "a",
32
+ turnId: "turn_1",
33
+ toolName: "t",
34
+ args: {},
35
+ requestedAt: "2026-01-01T00:00:00.000Z",
36
+ });
37
+ const second = gate.wait({
38
+ sessionId: "s1",
39
+ callId: "a2",
40
+ turnId: "turn_1",
41
+ toolName: "t2",
42
+ args: {},
43
+ requestedAt: "2026-01-01T00:00:00.000Z",
44
+ });
45
+ gate.cancelSession("s1", "deny");
46
+ await expect(first).resolves.toEqual({ decision: "deny" });
47
+ await expect(second).resolves.toEqual({ decision: "deny" });
48
+ });
49
+ });