@afokapu/atdd-bun 0.10.5 → 0.10.7

package/README.md

@@ -171,6 +171,10 @@ From 0.10.5 only coordinators and drivers use the board: a writer or reviewer ta
 launch prompt and returns its answer as output, which the driver posts. `atdd-bun chat wait` takes several
 topics and `--skip heartbeat`; reviewers answer in JSON.
 
+From 0.10.7 only a browser spec (a `*.e2e.*` file, or a Playwright file bound by `// Train:`/`// Journey:` or a
+`test:train:`/`test:journey:` URN) is a journey spec. A `bun:test` file may carry the `Acceptance:` line
+traceability needs, including a test proving `acc:train:…` and the journey test `atdd-bun.topology.e2e-location` requires.
+
 Every key is optional; these are the defaults:
 
 ```yaml
@@ -181,11 +185,11 @@ delivery:
   # board: { url: http://127.0.0.1:2586 }   # opt-in: agents talk through a local board; absent, there is none
   independence: different-model      # a reviewer's model wrote none of the work it reviews; or fresh-process
   stages:                            # models in preference order: the first, then recorded fallbacks
-    plan:     { writer: [codex, claude-opus], reviewer: [glm, claude-opus, codex] }
-    red:      { writer: [glm, claude-sonnet, claude-opus, codex] }
-    green:    { writer: [glm, claude-sonnet, claude-opus, codex] }
-    refactor: { writer: [glm, claude-sonnet, claude-opus, codex] }
-    final:    { reviewer: [codex, glm, claude-opus] }
+    plan:     { writer: [codex, claude-opus, kimi], reviewer: [glm, claude-opus, codex, kimi] }
+    red:      { writer: [glm, claude-sonnet, deepseek-flash, claude-opus, codex, kimi] }
+    green:    { writer: [glm, claude-sonnet, deepseek-flash, claude-opus, codex, kimi] }
+    refactor: { writer: [glm, claude-sonnet, deepseek-flash, claude-opus, codex, kimi] }
+    final:    { reviewer: [codex, glm, claude-opus, kimi] }
   fallback: { after_failures: 3, within_minutes: 10, when_exhausted: block }   # or wait
   commands: {}                       # per model: { author: "...", review: "..." } overriding delivery.operating-model's defaults
 ```

package/conventions/delivery/delivery.board.convention.yaml

@@ -27,7 +27,7 @@ terms:
     atdd-bun chat post writes the header: from, to, participants and conversation always; kind (task, result, review-request, verdict, rebuttal, status, blocked, heartbeat), stage, reply-to, repo, branch, worktree, goal, base and sha when they apply. The body carries the whole instruction or answer, however long: the thread is the history. A heartbeat carries nothing else; anything that needs an answer goes as blocked, never inside a heartbeat.
 - term_id: commands
   text: >-
-    Post with the body on standard input; read what is on a topic; wait for the next message addressed to you on one or several topics (comma-separated), skipping kinds you do not act on (--skip heartbeat), bounded by --timeout so a tool call ends (exit 2: nothing yet, wait again with --since the last id); and, for a human, atdd-bun chat alone opens the board: every topic on the left, the selected conversation on the right, live.
+    Post with the body on standard input; read what is on a topic; wait for the next message addressed to you on one or several topics (comma-separated), skipping kinds you do not act on (--skip heartbeat), bounded by --timeout so a tool call ends (exit 2: nothing yet). wait remembers where each identity stopped on each set of topics (~/.atdd-board/cursors), so re-arming a watcher is always the same command; --since still overrides it; and, for a human, atdd-bun chat alone opens the board: every topic on the left, the selected conversation on the right, live.
   values:
     post: atdd-bun chat post <topic> --to <agent> --kind <kind> [--reply-to <id>] < message.md
     read: atdd-bun chat read <topic> [--mine]

package/conventions/delivery/delivery.config-schema.convention.yaml

@@ -21,11 +21,11 @@ content:
         root: docs/delivery/tranches
         independence: different-model
         stages:
-          plan:     { writer: [codex, claude-opus], reviewer: [glm, claude-opus, codex] }
-          red:      { writer: [glm, claude-sonnet, claude-opus, codex] }
-          green:    { writer: [glm, claude-sonnet, claude-opus, codex] }
-          refactor: { writer: [glm, claude-sonnet, claude-opus, codex] }
-          final:    { reviewer: [codex, glm, claude-opus] }
+          plan:     { writer: [codex, claude-opus, kimi], reviewer: [glm, claude-opus, codex, kimi] }
+          red:      { writer: [glm, claude-sonnet, deepseek-flash, claude-opus, codex, kimi] }
+          green:    { writer: [glm, claude-sonnet, deepseek-flash, claude-opus, codex, kimi] }
+          refactor: { writer: [glm, claude-sonnet, deepseek-flash, claude-opus, codex, kimi] }
+          final:    { reviewer: [codex, glm, claude-opus, kimi] }
 
     independence defaults to different-model. A stage may name a writer, a reviewer or both. The 0.9
     names plan_review, test_review, code_review and final_review are still read, as plan, red, refactor

package/conventions/delivery/delivery.operating-model.convention.yaml

@@ -29,7 +29,7 @@ terms:
     One line each, for the coordinator to wait on: PROGRAM_EVENT <tranche> <STAGE <stage> <sha> | WORKER_START <role> <model> <sha> | WORKER_END <role> <model> <verdict> | FALLBACK <role> <from>→<to> <reason> | PR_OPENED <url> | MERGED <sha> | BLOCKED <reason> | HEARTBEAT>.
 - term_id: commands
   text: >-
-    The default headless command per model, overridden per model under delivery.commands.<model>.author (the writer's command) and .review. {prompt} and {worktree} are substituted. Claude runs take project settings only, so a user's own allowances cannot widen a reviewer. glm runs through Pi, which has no permission system: as a reviewer it gets the read tool alone (the driver puts the diff in its prompt and records its output), and as a writer it runs only inside a container or sandbox limited to the worktree.
+    The default headless command per model, overridden per model under delivery.commands.<model>.author (the writer's command) and .review. {prompt} and {worktree} are substituted. Claude runs take project settings only, so a user's own allowances cannot widen a reviewer. glm, kimi (Moonshot kimi-k3, a replacement for codex and claude-opus) and deepseek-flash (a smaller coding model beside claude-sonnet) run through Pi, which has no permission system; kimi and deepseek-flash take their keys from MOONSHOT_API_KEY and DEEPSEEK_API_KEY, or point delivery.commands at a wrapper that provides them. Through Pi, as a reviewer it gets the read tool alone (the driver puts the diff in its prompt and records its output), and as a writer it runs only inside a container or sandbox limited to the worktree.
   values:
     codex:
       author: codex exec --cd {worktree} --sandbox workspace-write "{prompt}"
@@ -43,6 +43,12 @@ terms:
     glm:
       author: cd {worktree} && pi -p --no-session --no-extensions --no-skills --no-context-files --tools read,bash,edit,write "{prompt}"
       review: cd {worktree} && pi -p --no-session --no-extensions --no-skills --no-context-files --tools read "{prompt}"
+    kimi:
+      author: cd {worktree} && pi --provider moonshotai --model kimi-k3 -p --no-session --no-extensions --no-skills --no-context-files --tools read,bash,edit,write "{prompt}"
+      review: cd {worktree} && pi --provider moonshotai --model kimi-k3 -p --no-session --no-extensions --no-skills --no-context-files --tools read "{prompt}"
+    deepseek-flash:
+      author: cd {worktree} && pi --provider deepseek --model deepseek-flash -p --no-session --no-extensions --no-skills --no-context-files --tools read,bash,edit,write "{prompt}"
+      review: cd {worktree} && pi --provider deepseek --model deepseek-flash -p --no-session --no-extensions --no-skills --no-context-files --tools read "{prompt}"
 - term_id: example_record
   text: >-
     A tranche in progress: the plan written by codex and approved by glm; red, green and refactor written by glm; the final review requesting changes. Only the finding's outcome is ever added to an earlier entry.

package/conventions/tester.htmx/tester.htmx.journey-no-acceptance-marker.convention.yaml

@@ -3,7 +3,7 @@ rule_id: tester.htmx.journey-no-acceptance-marker
 kind: rule
 status: active
 name: A journey spec carries no Acceptance or WMBT header
-statement: A journey spec MUST NOT carry `// Acceptance:` or `// WMBT:` header lines. Bun/htmx realization of `tester.vite.journey-no-acceptance-marker`,
+statement: A journey spec, meaning a browser spec (a `*.e2e.*` file, or a Playwright file bound to a train or journey), MUST NOT carry `// Acceptance:` or `// WMBT:` header lines; a bun test file is never a journey spec. Bun/htmx realization of `tester.vite.journey-no-acceptance-marker`,
   strict.
 terms:
 - term_id: browser_spec

package/detectors/htmx_e2e_detector/checks/_e2e.mjs

@@ -65,7 +65,10 @@ export function collectSpecs(roots, excludes) {
   return [...out.values()];
 }
 
-export const isJourneySpec = (spec) => spec.e2eNamed || spec.journeyMarked;
+// A journey spec is a browser spec: a *.e2e.* file, or a Playwright file whose header binds a train or journey. A bun:test
+// file is never one, so a Bun test proving acc:train:… (or the journey test atdd-bun.topology.e2e-location requires) may
+// carry the Acceptance: line traceability needs.
+export const isJourneySpec = (spec) => spec.e2eNamed || (spec.playwright && spec.journeyMarked);
 
 /** The plan subjects a spec can bind to, read with the package's own loader. */
 export async function planOf(root) {

package/integrity.json

@@ -1,9 +1,9 @@
 {
-  "version": "0.10.5",
+  "version": "0.10.7",
   "files": {
     "HOOK_AUDIT.md": "5329d840db37671b1918f688ead26865473b87db73dbc75f7c8b2a8bbe8d6d43",
     "PLANNER_PORT.md": "fb5935bac8b7ac18994de21e43ace3a5ef8cd55f85b0e3349fca261280054f11",
-    "README.md": "b3d92666940ed0801e52bd8c4fd0f2942b841a52f4eceb543776d867117e2f75",
+    "README.md": "54f6959a5fe0a83ff7ba07856f5385eff0bff0607bfe7b2287897b2d8378825e",
     "bunfig.toml": "b9fc65eca9014c5179380259d70e76385a6a80788fa9a2df5fb3eaa5554fd2fe",
     "conventions/atdd-bun.planner/atdd-bun.planner.acceptance-identity.convention.yaml": "81c5c773d5ee15e8d99a2c237aa9846b110b2533df45ebf407d0e3ec6ed3dd03",
     "conventions/atdd-bun.planner/atdd-bun.planner.identity-required.convention.yaml": "e96d7c1455d0c221072d82e7f2b55da1c4f9ceb17718eb6ed96affdeef37675b",
@@ -92,13 +92,13 @@
     "conventions/coder.htmx/coder.htmx.verb-endpoint-same-origin.convention.yaml": "2128d26153e2d1f19c65a1e1665d4f82c78395daedb336d6ef1256e82b74aedb",
     "conventions/coder.htmx/coder.htmx.verb-mutation-signals-progress.convention.yaml": "ae85c5ea01a09a3c41b939988088a8e5da91848d2710a1f0a1e0c86308638ac9",
     "conventions/delivery/delivery.approved-sha-resolves.convention.yaml": "12eadbc75b081e9469b9a0a7dbf3fdd30c8b34f66e647efa8bf2528119e61547",
-    "conventions/delivery/delivery.board.convention.yaml": "c61df91a5fa95ceb77be42d0571a0f904f76ed2324cd61ae17e0d87caf42cec9",
-    "conventions/delivery/delivery.config-schema.convention.yaml": "455001b7ac12824e27b1f1dd13ae8c09bfa140957e118ebd1229779ea0808d4f",
+    "conventions/delivery/delivery.board.convention.yaml": "51ff4c16dddf20504f34accc5ddcd150108fab70e4e6f5bdb36cec71d0c905a8",
+    "conventions/delivery/delivery.config-schema.convention.yaml": "4c7b449074a8bd9f89b70c702ab3e0df229105ddb4458772f894f0bf37813e9f",
     "conventions/delivery/delivery.evidence-schema.convention.yaml": "9c6a22495dfb3e2fcbca700392866c3b5660084b1f5594230eb320b499d5e5cb",
     "conventions/delivery/delivery.findings-resolved.convention.yaml": "f6574e93a9a2d4bdd6d6f5645720b1a647fc0048d95f7675fe783d2ddb62c05b",
     "conventions/delivery/delivery.merge-gate.convention.yaml": "48745ec23625b68be9b8fe2c6dcdcba88b2b38924f6593abf6987391ff0ee243",
     "conventions/delivery/delivery.model-allowed.convention.yaml": "1071a727f61bacd21d70921885826974234ef0a58c0a814454b9375a0432b0fe",
-    "conventions/delivery/delivery.operating-model.convention.yaml": "e66f75787518954cd830fff544d40ac7ddecdf8cf58aa1d73a76a30beeff3fe9",
+    "conventions/delivery/delivery.operating-model.convention.yaml": "7982157d957938ade878a114d3f5c2bdaa8f02e52ac21614634d9af4be9692f1",
     "conventions/delivery/delivery.review.convention.yaml": "b033579c2d90fcb098d325dcc07c7b084e048cca43fbe8e28159ec08a412018a",
     "conventions/delivery/delivery.reviewer-independent.convention.yaml": "8c436d66ef5b1f6d6bc5dd55a235caca2490020e88665eecadb2ca41cd4149c5",
     "conventions/delivery/delivery.stages-complete.convention.yaml": "0e5dd8057ad51eff7906e3424a9686a81be1814447afb9a30acab43d39e49b4c",
@@ -150,7 +150,7 @@
     "conventions/tester.htmx/tester.htmx.fragment-asserts-returned-markup.convention.yaml": "b8de8e893b582d8ec4b7c24acfa16da0eccd14e6e58e92baf9e9c10c9749ea0c",
     "conventions/tester.htmx/tester.htmx.journey-binding-header.convention.yaml": "7a9f8655cccdd12ab7e2d49af128c9f9572ad441823ca6e32cbae26b26b2d714",
     "conventions/tester.htmx/tester.htmx.journey-layer-assembly.convention.yaml": "86948c4b1e928eef07329afeec73660ea51878e3baa58350f61a8f81c20e9fc3",
-    "conventions/tester.htmx/tester.htmx.journey-no-acceptance-marker.convention.yaml": "548588cb2722fb906034af7cc98bcce91000afb1cfd20875a96f143f71e43a18",
+    "conventions/tester.htmx/tester.htmx.journey-no-acceptance-marker.convention.yaml": "f0d96af21e057e8ae02bb0d3c3f6e10c136e5b3eb19d94a7d0f034bc0b0a87e0",
     "conventions/tester.htmx/tester.htmx.journey-urn-format.convention.yaml": "6771d984ddb6fe564a6803e971e934242d0a53bd8711826e66d4d0c3283f8aed",
     "conventions/tester.htmx/tester.htmx.presentation-smoke-coverage.convention.yaml": "b18565f55f9f92f7b292bae8231c2343e72f910dc9e228b307db318eb62488cc",
     "conventions/tester.htmx/tester.htmx.responsive-harness.convention.yaml": "cec312935ab592c2c7a5c6a94623a02df771a9e2ca2035b721ef1006c4800fdf",
@@ -675,7 +675,7 @@
     "detectors/delivery_evidence/fixtures/dirty/docs/delivery/tranches/orphan/README.md": "7d480b1fff6546615e55d866784bc2267d06d0221dfde8d0ba0353fa1ea0db9a",
     "detectors/delivery_evidence/fixtures/dirty/docs/delivery/tranches/ui/evidence.yaml": "777b4f7921e302dfce79cd764bba8b44091ac9c7ee04a00f5dbb663342ec68a0",
     "detectors/htmx_e2e_detector/atdd.implementation.yaml": "454bb123039d57c992d27d9274e20fa3452ac41a237eb55471be62a4bcfebf06",
-    "detectors/htmx_e2e_detector/checks/_e2e.mjs": "4e3605500ac232603083790c3984de5c828028c60eef366b26aba18d38fce30b",
+    "detectors/htmx_e2e_detector/checks/_e2e.mjs": "efec6c8e3dc8b95bf3acd826397ecdedac64c06d367791a4017a272bc0660896",
     "detectors/htmx_e2e_detector/checks/_map.json": "94dec08ff4764cd6a3dba139f12a7c984292e95a9354b0cdaa0ce9d4180e2c89",
     "detectors/htmx_e2e_detector/checks/a11y_harness.mjs": "d66e2a6da3c3820db72fdaddcd53efb9c0e39d03bc559a59b800923d5c944700",
     "detectors/htmx_e2e_detector/checks/covered_train_is_routed.mjs": "c7a8db0e932ba96489477fb5a8589cd9eba868db6fbaaa8012802a4da78fa0f8",
@@ -1192,10 +1192,10 @@
     "relationships.yaml": "97da2db334e3c8cd4f634ee8d12fe426aabc0f5c37508d02450ce8beca100f04",
     "src/agent.ts": "65e4579adb5a54977f557f8c783aba0e2a2a1d5cf3cea9f017b67ffe598b2f65",
     "src/board-ui.ts": "16a843705552f8f18e0e328e90e4a2a9148d771ca69f2708fe4a91302947dfad",
-    "src/board.ts": "dede8d3059165a1a0fd051420d83ef91ff9a5fffa57d1ff6c44f31defa393bfa",
+    "src/board.ts": "115c6cff12fb2b20e6471013161aef020a65331ecb97fa323e9b08f55053149f",
     "src/ci.ts": "67e54d2cbfc44a9e42837d5e750af526d6255c539d3c452a4d5fca3226c0d9ac",
     "src/cli.ts": "e45f2666334503a91eb65b1b5ee4dd8acab6b99f62d333e2ee2b65963db6358b",
-    "src/delivery.ts": "3fd36b6f7dbe9b5682241cff16362618ac0a85c0ac93a238141ebabdbe402e52",
+    "src/delivery.ts": "9a4500ae949faeef7bcfda114b7b09d4ac2b2cbb1218e68016829e0b6892847b",
     "src/docs-capability.ts": "115cf27049a5cf19c133bac5ec237072be6f969a59a2ef6767e054108a21dc5c",
     "src/enforce.ts": "41a6f8c058a2a034996c817f224191403cc66ccc4ed4e04bbefc430014c783d8",
     "src/hooks.ts": "9779eb48f9d0cb6234c33f6616f98dde6583d2ee03ac3c19f3ace2d156af26d0",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@afokapu/atdd-bun",
-  "version": "0.10.5",
+  "version": "0.10.7",
   "repository": {
     "type": "git",
     "url": "git+https://github.com/afokapu/atdd-bun.git"

package/src/board.ts

@@ -1,5 +1,6 @@
 import { createHash } from "node:crypto";
-import { existsSync } from "node:fs";
+import { existsSync, mkdirSync, readFileSync, writeFileSync } from "node:fs";
+import { homedir } from "node:os";
 import { readFile } from "node:fs/promises";
 import { join } from "node:path";
 import { runUi } from "./board-ui";
@@ -107,17 +108,26 @@ export async function read(url: string, topics: string[], since = "all"): Promis
  * It polls rather than streams: a live stream can drop a message published while it connects. ntfy reads `since` (a
  * message id) as that message's time on every topic, so one cursor serves several topics. Returns null after
  * `timeoutSeconds`, so an agent's tool call stays bounded. */
-export async function waitFor(url: string, topics: string[], agent: string, since = "all", timeoutSeconds = 0, skip: string[] = [], intervalMs = 500): Promise<Message | null> {
+export async function waitFor(url: string, topics: string[], agent: string, since = "all", timeoutSeconds = 0, skip: string[] = [], intervalMs = 500, seen: (id: string) => void = () => {}): Promise<Message | null> {
   const deadline = timeoutSeconds > 0 ? Date.now() + timeoutSeconds * 1000 : Infinity;
   for (;;) {
     let messages: Message[] = [];
     try { messages = await read(url, topics, since); } catch { /* the board is briefly unavailable: keep waiting */ }
-    for (const message of messages) { since = message.id; if (addressedTo(message, agent) && !skip.includes(message.header.kind ?? "")) return message; }
+    for (const message of messages) { since = message.id; seen(since); if (addressedTo(message, agent) && !skip.includes(message.header.kind ?? "")) return message; }
     if (Date.now() >= deadline) return null;
     await Bun.sleep(intervalMs);
   }
 }
 
+/** Where `wait` keeps its place, per identity and topic set, so restarting a watcher is always the same command.
+ * Best effort: where the file cannot be read or written, wait behaves as if it had no saved place. */
+function cursorFile(agent: string, topics: string[], env: Record<string, string | undefined>): string {
+  const key = `${agent}__${[...topics].sort().join("+")}`.replace(/[^A-Za-z0-9@+_.-]/g, "_");
+  return join(env.ATDD_BOARD_STATE || join(homedir(), ".atdd-board", "cursors"), key);
+}
+const loadCursor = (file: string) => { try { return readFileSync(file, "utf8").trim() || undefined; } catch { return undefined; } };
+const saveCursor = (file: string, id: string) => { try { mkdirSync(join(file, ".."), { recursive: true }); writeFileSync(file, `${id}\n`); } catch { /* best effort */ } };
+
 export function format(message: Message, withTopic = false): string {
   const time = new Date(message.time * 1000).toTimeString().slice(0, 8), h = message.header;
   const route = `${h.from ?? "?"} -> ${[h.to ?? []].flat().join(", ") || "?"}`, tags = [h.kind, h.stage].filter(Boolean).join(" · ");
@@ -141,8 +151,10 @@ export async function chat(args: string[], env: Record<string, string | undefine
     const me = agent();
     topics.forEach(topic => checkTopic(topic, env));
     if (command === "wait") {
-      const message = await waitFor(url, topics, me, flag("since") || "all", Number(flag("timeout") ?? 0), list(flag("skip")));
-      if (!message) { console.error(`no message for ${me} on ${topics.join(", ")} yet; wait again with --since to continue`); return 2; }
+      // Without --since, continue from where this identity last stopped on these topics; either way, remember the place.
+      const file = cursorFile(me, topics, env);
+      const message = await waitFor(url, topics, me, flag("since") || loadCursor(file) || "all", Number(flag("timeout") ?? 0), list(flag("skip")), 500, id => saveCursor(file, id));
+      if (!message) { console.error(`no message for ${me} on ${topics.join(", ")} yet; run the same wait again to continue`); return 2; }
       console.log(format(message, topics.length > 1));
       return 0;
     }

package/src/delivery.ts

@@ -41,11 +41,11 @@ export type DeliveryPolicy = {
 /** The default operating model: two reviews, the plan and the whole change; the stages between them are written and
  * held by their deterministic gates. Lists are preference orders; a later model is used only as a recorded fallback. */
 const DEFAULT_STAGES: Partial<Record<Stage, { writer?: string[]; reviewer?: string[] }>> = {
-  plan: { writer: ["codex", "claude-opus"], reviewer: ["glm", "claude-opus", "codex"] },
-  red: { writer: ["glm", "claude-sonnet", "claude-opus", "codex"] },
-  green: { writer: ["glm", "claude-sonnet", "claude-opus", "codex"] },
-  refactor: { writer: ["glm", "claude-sonnet", "claude-opus", "codex"] },
-  final: { reviewer: ["codex", "glm", "claude-opus"] },
+  plan: { writer: ["codex", "claude-opus", "kimi"], reviewer: ["glm", "claude-opus", "codex", "kimi"] },
+  red: { writer: ["glm", "claude-sonnet", "deepseek-flash", "claude-opus", "codex", "kimi"] },
+  green: { writer: ["glm", "claude-sonnet", "deepseek-flash", "claude-opus", "codex", "kimi"] },
+  refactor: { writer: ["glm", "claude-sonnet", "deepseek-flash", "claude-opus", "codex", "kimi"] },
+  final: { reviewer: ["codex", "glm", "claude-opus", "kimi"] },
 };
 /** The stage a record or config name refers to: a current name, or a legacy review stage's. */
 export const stageOf = (name: string): Stage | undefined => (STAGES as readonly string[]).includes(name) ? name as Stage : LEGACY_STAGES[name];