@gobing-ai/knowledge-kit 0.0.32 → 0.0.33

package/plugins/ingestions/last30days-ingest/dist/index.js

@@ -14642,7 +14642,7 @@ var init_zod = __esm(() => {
   init_external();
 });
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/errors.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/errors.js
 var ErrorCode;
 var init_errors3 = __esm(() => {
   ErrorCode = {
@@ -14653,7 +14653,7 @@ var init_errors3 = __esm(() => {
   };
 });
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/api-response.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/api-response.js
 var API_ERROR_CODES, ERROR_CODE_TO_HTTP, CLIENT_SAFE_CODES;
 var init_api_response = __esm(() => {
   init_errors3();
@@ -14676,13 +14676,13 @@ var init_api_response = __esm(() => {
   CLIENT_SAFE_CODES = new Set([ErrorCode.NotFound, ErrorCode.Validation, ErrorCode.Conflict]);
 });
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/cursor.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/cursor.js
 var init_cursor = () => {};
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/event-severity.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/event-severity.js
 var init_event_severity = () => {};
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/output.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/output.js
 function processStream(name) {
   const proc = globalThis.process;
   const stream = proc?.[name];
@@ -14700,7 +14700,7 @@ function echoError(message, target = defaultStderrTarget ?? processStream("stder
 }
 var defaultStderrTarget;
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/index.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/index.js
 var init_dist = __esm(() => {
   init_api_response();
   init_cursor();
@@ -21698,7 +21698,7 @@ var init_dist2 = __esm(() => {
   $visitAsync = visit.visitAsync;
 });
 
-// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.15/node_modules/@gobing-ai/ts-runtime/dist/config.js
+// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.18+ea28d47622d670ff/node_modules/@gobing-ai/ts-runtime/dist/config.js
 var configSchema;
 var init_config = __esm(() => {
   init_dist2();
@@ -21722,7 +21722,7 @@ var init_config = __esm(() => {
   });
 });
 
-// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.15/node_modules/@gobing-ai/ts-runtime/dist/file-system-node.js
+// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.18+ea28d47622d670ff/node_modules/@gobing-ai/ts-runtime/dist/file-system-node.js
 import { appendFileSync, cpSync, createReadStream, createWriteStream, existsSync, mkdirSync, readdirSync, readFileSync, realpathSync, renameSync, rmSync, statSync, writeFileSync } from "fs";
 import { dirname, resolve as resolvePath } from "path";
 function createNodeFileSystem(root, paths) {
@@ -21898,18 +21898,19 @@ var PluginManifestSchema = exports_external.looseObject({
   name: exports_external.string().min(1),
   kind: PluginKindSchema,
   entry: exports_external.string().min(1),
+  check: exports_external.string().min(1).optional(),
   version: exports_external.string().optional(),
   description: exports_external.string().optional(),
   operations: exports_external.record(exports_external.string(), PluginOperationSchema).optional()
 });
-// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.15/node_modules/@gobing-ai/ts-runtime/dist/index.js
+// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.18+ea28d47622d670ff/node_modules/@gobing-ai/ts-runtime/dist/index.js
 init_file_system_node();
 init_config();
 
-// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.15/node_modules/@gobing-ai/ts-runtime/dist/fs.js
+// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.18+ea28d47622d670ff/node_modules/@gobing-ai/ts-runtime/dist/fs.js
 init_file_system_node();
 
-// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.15/node_modules/@gobing-ai/ts-runtime/dist/path.js
+// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.18+ea28d47622d670ff/node_modules/@gobing-ai/ts-runtime/dist/path.js
 function normalizeSeparators(path) {
   return path.replaceAll("\\", "/");
 }
@@ -21943,7 +21944,7 @@ function dirnamePath(path) {
   return normalized.slice(0, index);
 }
 
-// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.15/node_modules/@gobing-ai/ts-runtime/dist/fs.js
+// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.18+ea28d47622d670ff/node_modules/@gobing-ai/ts-runtime/dist/fs.js
 async function ensureDirForFile(path, fs = createNodeFileSystem()) {
   await fs.ensureDir(dirnamePath(path));
 }

package/plugins/ingestions/web-search/dist/index.js

@@ -14642,7 +14642,7 @@ var init_zod = __esm(() => {
   init_external();
 });
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/errors.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/errors.js
 var ErrorCode;
 var init_errors3 = __esm(() => {
   ErrorCode = {
@@ -14653,7 +14653,7 @@ var init_errors3 = __esm(() => {
   };
 });
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/api-response.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/api-response.js
 var API_ERROR_CODES, ERROR_CODE_TO_HTTP, CLIENT_SAFE_CODES;
 var init_api_response = __esm(() => {
   init_errors3();
@@ -14676,13 +14676,13 @@ var init_api_response = __esm(() => {
   CLIENT_SAFE_CODES = new Set([ErrorCode.NotFound, ErrorCode.Validation, ErrorCode.Conflict]);
 });
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/cursor.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/cursor.js
 var init_cursor = () => {};
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/event-severity.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/event-severity.js
 var init_event_severity = () => {};
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/output.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/output.js
 function processStream(name) {
   const proc = globalThis.process;
   const stream = proc?.[name];
@@ -14700,7 +14700,7 @@ function echoError(message, target = defaultStderrTarget ?? processStream("stder
 }
 var defaultStderrTarget;
 
-// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.15/node_modules/@gobing-ai/ts-utils/dist/index.js
+// ../../node_modules/.bun/@gobing-ai+ts-utils@0.5.18/node_modules/@gobing-ai/ts-utils/dist/index.js
 var init_dist = __esm(() => {
   init_api_response();
   init_cursor();
@@ -21698,7 +21698,7 @@ var init_dist2 = __esm(() => {
   $visitAsync = visit.visitAsync;
 });
 
-// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.15/node_modules/@gobing-ai/ts-runtime/dist/config.js
+// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.18+ea28d47622d670ff/node_modules/@gobing-ai/ts-runtime/dist/config.js
 var configSchema;
 var init_config = __esm(() => {
   init_dist2();
@@ -21722,7 +21722,7 @@ var init_config = __esm(() => {
   });
 });
 
-// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.15/node_modules/@gobing-ai/ts-runtime/dist/file-system-node.js
+// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.18+ea28d47622d670ff/node_modules/@gobing-ai/ts-runtime/dist/file-system-node.js
 import { appendFileSync, cpSync, createReadStream, createWriteStream, existsSync, mkdirSync, readdirSync, readFileSync, realpathSync, renameSync, rmSync, statSync, writeFileSync } from "fs";
 import { dirname, resolve as resolvePath } from "path";
 function createNodeFileSystem(root, paths) {
@@ -21897,11 +21897,12 @@ var PluginManifestSchema = exports_external.looseObject({
   name: exports_external.string().min(1),
   kind: PluginKindSchema,
   entry: exports_external.string().min(1),
+  check: exports_external.string().min(1).optional(),
   version: exports_external.string().optional(),
   description: exports_external.string().optional(),
   operations: exports_external.record(exports_external.string(), PluginOperationSchema).optional()
 });
-// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.15/node_modules/@gobing-ai/ts-runtime/dist/index.js
+// ../../node_modules/.bun/@gobing-ai+ts-runtime@0.5.18+ea28d47622d670ff/node_modules/@gobing-ai/ts-runtime/dist/index.js
 init_file_system_node();
 init_config();
 
@@ -24036,7 +24037,7 @@ var ConfigError = class extends Error {
   }
 };
 
-// ../../node_modules/.bun/@gobing-ai+ts-infra@0.5.15+c7cc99588d9d7f09/node_modules/@gobing-ai/ts-infra/dist/logger.js
+// ../../node_modules/.bun/@gobing-ai+ts-infra@0.5.18+54f059b6897bbcdb/node_modules/@gobing-ai/ts-infra/dist/logger.js
 var ROOT_CATEGORY = "app";
 function toLogTapeLevel(level) {
   return level === "warn" ? "warning" : level;

package/plugins/kk/README.md

@@ -11,6 +11,7 @@ It is a Claude Code / Superskill plugin, **not** a kk-core product plugin.
 | `hooks/` | Hook definitions |
 | `rules/` | Agent rules |
 | `workflows/` | Design-time product YAML SSOT (ADR-009). Runtime dest is `$HOME/.config/kk/workflows` via `/workflow-run` |
+| `schedules/` | launchd/cron **templates** for the shipped workflows (`kk schedule render` prints them) |
 
 `agents/` is reserved for thin wrappers of fat skills. It is empty: `judge-tech` /
 `judge-tone` / `judge-compliance` were retired (2026-08-19). Rubric selection lives on
@@ -26,3 +27,40 @@ Capability files **omit a leading `kk-`**. `superskill install` prefixes the plu
 `skills/content-judge` installs as `kk:content-judge` (not `kk:kk-content-judge`). Same for `topic`,
 `storm-research`, `itc-generating`, `explain-things` (via `/tell-me`), and `/workflow-run`. Product workflow YAML (`kk-storm-research.yaml`)
 keeps its existing name — that is a workflow stem, not an installable agent capability.
+
+## Schedule
+
+`spur` **has no scheduler**, so kk ships templates and `kk schedule render` prints the artifact for
+the platform — **installing it is an operator action** (kk never writes a crontab or a
+`LaunchAgents` plist):
+
+```bash
+# launchd (macOS): every 6 h at 04:00/10:00/16:00/22:00 local (StartCalendarInterval), one shared log
+# 1. the log directory must exist — launchd opens StandardOutPath but never creates its parent
+mkdir -p "$PWD/.works/kk-collect"
+# 2. render + bootstrap (the plist carries the PATH launchd's default lacks)
+kk schedule render kk-collect --platform launchd --repo "$PWD" \
+  > ~/Library/LaunchAgents/ai.gobing.kk-collect.plist
+launchctl bootstrap gui/$(id -u) ~/Library/LaunchAgents/ai.gobing.kk-collect.plist
+
+# cron: the same 6-hourly cadence (the rendered line carries its own mkdir + PATH).
+# Merge, never replace: `| crontab -` discards every other job in the operator's crontab.
+crontab -l > /tmp/kk-collect.cron 2>/dev/null || :
+kk schedule render kk-collect --platform cron --repo "$PWD" >> /tmp/kk-collect.cron
+crontab /tmp/kk-collect.cron
+```
+
+Both rendered artifacts set `PATH` themselves (`{{home}}/.bun/bin`, `/opt/homebrew/bin`,
+`/usr/local/bin`, then the system dirs). That is what the workflow's **children** need — `kk`, `jq` and
+the plugins are all `#!/usr/bin/env bun` or PATH-resolved, and `launchd`/`cron` start with a minimal
+`PATH`. It is **not** what resolves the scheduler's own command: launchd looks up a *relative*
+`ProgramArguments[0]` with its fixed `_PATH_STDPATH` before the job environment exists, so the plist
+also carries an absolute `Program` (`{{home}}/.bun/bin/spur`) while keeping `ProgramArguments` as the
+frozen argv. Set both to wherever `spur`/`bun` actually live, and `mkdir -p` the log directory first
+(cron's `>>` aborts before the command runs when its directory is missing).
+
+`kk-collect` collects every ingestion source into the kk store and never publishes. Two ticks that
+overlap are **one** collection: the first state takes `kk store lock acquire --name collect` and a
+tick that finds the claim inside its TTL exits 0 (`skipped`) without collecting. Before relying on
+the schedule, check the store is reachable from the scheduled checkout (`kk store migrate`), and
+`tail -f "$PWD/.works/kk-collect/launchd.log"` after the first launch.

package/plugins/kk/decisions/extract/ai-news.yaml

@@ -0,0 +1,21 @@
+# ai-news topic profile (task 0210 R1, design kk-collect-and-daily-voice2.md §4).
+#
+# The episode pool for the daily AI-news workflow: every labelled item in the
+# launch/industry categories with a substantive body and real audience signal,
+# most time-sensitive first. `kk store extract --topic ai-news` compiles this into
+# one parameterized query over `items` + `labels` — a new topic is a new file here,
+# never a code change.
+#
+# Keys are decision ids from the sibling catalog (plugins/kk/decisions/kk-items.yaml);
+# an unknown key or decision id fails loud with `kk.store.profile.invalid`.
+# `limit` is omitted: the compiler's episode-cap default (20) applies unless
+# `kk store extract --limit` overrides it.
+topic: ai-news
+where:
+  kk-item-category: { in: [ai-news, ai-industry] }
+  # quality 2, not 3: real ingest bodies (last30days/aihot summaries) are stub-length, so the
+  # heuristic labeler tops out at 1-2 and a gte-3 gate empties every real pool (task 0212 dogfood
+  # finding). Raise this back when a full-text ingest source lands.
+  kk-item-quality: { gte: 2 }
+  kk-item-importance: { gte: 3 }
+order: [kk-item-urgency desc, kk-item-impact desc, first_seen_at desc]

package/plugins/kk/decisions/kk-items.yaml

@@ -0,0 +1,113 @@
+# kk item-label decision catalog (task 0209, design §4).
+#
+# One decision per item label `kk store label` writes: the MECE category plus the
+# four 0–5 QC dimensions. The rubric texts and the choice vocabulary mirror
+# episode-plan-gen's keyword heuristic (`classifyCategory` / `scoreDimensions`), so a
+# model answer and a heuristic answer are directly comparable rows in `labels`.
+#
+# Resolution: registered through `decisions.paths` in `.spur/config.yaml`
+# (project-relative `plugins/kk/decisions`). No kk workflow reads this catalog as a
+# `decide` action; `kk store label` calls `spur decision run` and degrades to the
+# heuristic when the backend does not answer.
+$schema: "@gobing-ai/spur/schemas/decision-catalog.schema.json"
+version: 1
+
+defaults:
+    # ADR-125 value, in lockstep with config/decisions/task-pipeline.yaml. A model answer
+    # below this resolves to the decision's declared fallback with source: default.
+    minConfidence: 0.8
+
+decisions:
+    kk-item-category:
+        type: choice
+        description: Which MECE category does this collected item belong to?
+        parameters:
+            title:
+                type: string
+                description: Item title (the heuristic's first keyword surface).
+            source:
+                type: string
+                description: Ingest plugin that produced the item (e.g. aihot, rss, polymarket).
+        criteria:
+            ai-news: Launch, release, announcement or product news about AI — the company/lab keyword surface.
+            ai-technical: Engineering substance — APIs, SDKs, tutorials, benchmarks, architectures, inference.
+            ai-research: Research output — papers, alignment, evaluations, novel findings.
+            ai-industry: Company and market signals — funding, acquisitions, valuation, policy, regulation, litigation.
+            non-ai: No AI signal survives the keyword scan, or a deny rule (non-AI source, junk phrase, non-AI tag) wins.
+        fallback: non-ai
+
+    kk-item-quality:
+        type: score
+        description: How substantive is the item's body (length and sourcedness)?
+        parameters:
+            title:
+                type: string
+                description: Item title.
+            source:
+                type: string
+                description: Ingest plugin that produced the item.
+        criteria:
+            - Empty or unusable body — nothing to read.
+            - Headline echo — a stub body under 300 characters, source mention only.
+            - Around 1.5k characters — a short but substantive report.
+            - Around 3k characters — a fully developed report.
+            - Around 4.5k characters — a detailed report with context.
+            - Around 6k characters or more, or any substantive body carrying a verifiable source URL.
+        fallback: 2
+
+    kk-item-importance:
+        type: score
+        description: How much does the item matter to an AI audience (source tier plus relevance)?
+        parameters:
+            title:
+                type: string
+                description: Item title.
+            source:
+                type: string
+                description: Ingest plugin that produced the item.
+        criteria:
+            - No audience signal at all.
+            - Tail source or unknown origin — incidental aggregation.
+            - Tail source (reddit, youtube, github, polymarket) — worth a slot on a thin day.
+            - Ranked feed (rss, horizon) — a horizon-ranked story.
+            - Human-curated front page (aihot) — the editor's own pick.
+            - Any tier whose relevance score is 0.8 or higher — a signal-confirmed story.
+        fallback: 2
+
+    kk-item-urgency:
+        type: score
+        description: How time-sensitive is the item relative to the plan date?
+        parameters:
+            title:
+                type: string
+                description: Item title.
+            source:
+                type: string
+                description: Ingest plugin that produced the item.
+        criteria:
+            - No publication date and no curated-recency signal — urgency cannot be judged (the heuristic never files this level).
+            - Published more than 7 days before the plan date — stale, archive material.
+            - Published 3 to 7 days before the plan date — cooling (the heuristic never files this level).
+            - Published within the last 7 days, or an undated same-day curated source — the default.
+            - Published 1 to 2 days before the plan date — still breaking.
+            - Published on or after the plan date — same-day news.
+        fallback: 2
+
+    kk-item-impact:
+        type: score
+        description: How large is the item's measured engagement?
+        parameters:
+            title:
+                type: string
+                description: Item title.
+            source:
+                type: string
+                description: Ingest plugin that produced the item.
+        criteria:
+            - No engagement recorded at all — unmeasurable reach (the heuristic never files this level).
+            - Engagement magnitude 1 or less — a single mention.
+            - Engagement magnitude above 1 — the item is being shared.
+            - Engagement magnitude above 10 — a discussion thread.
+            - Engagement magnitude above 100 — a widely shared story.
+            - Engagement magnitude above 1000 — a breakout story.
+        fallback: 2

package/plugins/kk/plugin.json

@@ -1,5 +1,5 @@
 {
     "name": "kk",
-    "version": "0.0.32",
+    "version": "0.0.33",
     "description": "knowledge-kit Claude Code plugin: skills, commands, subagents, hooks, and rules"
 }

package/plugins/kk/schedules/cron.tmpl

@@ -0,0 +1,15 @@
+# kk-collect — standalone 6-hourly collection (04:00/10:00/16:00/22:00 local) into the kk store.
+#
+# Rendered by `kk schedule render kk-collect --platform cron --repo <checkout>`; kk never edits
+# the operator's crontab. Operator action (plugins/kk/README.md §Schedule) — MERGE, never replace:
+#
+#     crontab -l > /tmp/kk-collect.cron 2>/dev/null || :
+#     kk schedule render kk-collect --platform cron --repo "$PWD" >> /tmp/kk-collect.cron
+#     crontab /tmp/kk-collect.cron
+#
+# `kk store lock` is the single-flight guard, so an overlapping tick exits 0 without collecting.
+#
+# `mkdir -p` + an explicit PATH are load-bearing, not decoration: cron runs with a minimal PATH (no
+# `spur`/`kk`/`jq`/`bun`) and its `>>` redirect fails on a missing directory BEFORE the command runs,
+# so a bare line would produce no log and no tick. The README install steps mirror this.
+0 4,10,16,22 * * * cd {{repo}} && mkdir -p .works/kk-collect && PATH="{{home}}/.bun/bin:/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin" spur workflow run plugins/kk/workflows/kk-collect.yaml >> {{repo}}/.works/kk-collect/launchd.log 2>&1

package/plugins/kk/schedules/launchd.plist.tmpl

@@ -0,0 +1,77 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<!-- kk-collect — standalone 6-hourly collection (04:00/10:00/16:00/22:00 local) into the kk store.
+
+     Rendered by `kk schedule render kk-collect --platform launchd --repo <checkout>`; kk never
+     installs it. Operator action (plugins/kk/README.md §Schedule):
+
+         kk schedule render kk-collect --platform launchd --repo "$PWD" \
+           > ~/Library/LaunchAgents/ai.gobing.kk-collect.plist
+         launchctl bootstrap gui/$(id -u) ~/Library/LaunchAgents/ai.gobing.kk-collect.plist
+
+     StartCalendarInterval fires at 04:00/10:00/16:00/22:00 local wall-clock time (6 h cadence);
+     hours the machine is asleep through coalesce into one run at wake (launchd.plist(5)), where cron
+     would skip. The engine runs one node at a time and `kk store lock` is the single-flight guard,
+     so overlap is answered with `skipped`, never with two collectors. -->
+<plist version="1.0">
+<dict>
+    <key>Label</key>
+    <string>ai.gobing.kk-collect</string>
+    <!-- launchd resolves a RELATIVE ProgramArguments[0] with its own _PATH_STDPATH
+         (/usr/bin:/bin:/usr/sbin:/sbin) before the job environment exists, so a bare `spur` fails with
+         posix_spawn error 0x2 no matter what PATH this plist sets (man launchd.plist: ProgramArguments).
+         `Program` is the absolute executable launchd execs; ProgramArguments keeps R6's frozen argv. -->
+    <key>Program</key>
+    <string>{{home}}/.bun/bin/spur</string>
+    <key>ProgramArguments</key>
+    <array>
+        <string>spur</string>
+        <string>workflow</string>
+        <string>run</string>
+        <string>plugins/kk/workflows/kk-collect.yaml</string>
+    </array>
+    <key>WorkingDirectory</key>
+    <string>{{repo}}</string>
+    <key>StartCalendarInterval</key>
+    <array>
+        <dict>
+            <key>Hour</key>
+            <integer>4</integer>
+            <key>Minute</key>
+            <integer>0</integer>
+        </dict>
+        <dict>
+            <key>Hour</key>
+            <integer>10</integer>
+            <key>Minute</key>
+            <integer>0</integer>
+        </dict>
+        <dict>
+            <key>Hour</key>
+            <integer>16</integer>
+            <key>Minute</key>
+            <integer>0</integer>
+        </dict>
+        <dict>
+            <key>Hour</key>
+            <integer>22</integer>
+            <key>Minute</key>
+            <integer>0</integer>
+        </dict>
+    </array>
+    <!-- The job environment applies only AFTER exec, so it is not what resolves `spur` above —
+         it is what the workflow's own children need (`kk`, `jq`, `bun` are all `#!/usr/bin/env bun`
+         or PATH-resolved). -->
+    <key>EnvironmentVariables</key>
+    <dict>
+        <key>PATH</key>
+        <string>{{home}}/.bun/bin:/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin</string>
+    </dict>
+    <key>StandardOutPath</key>
+    <string>{{repo}}/.works/kk-collect/launchd.log</string>
+    <key>StandardErrorPath</key>
+    <string>{{repo}}/.works/kk-collect/launchd.log</string>
+    <key>RunAtLoad</key>
+    <false/>
+</dict>
+</plist>