@wonsukchoi/crondex 0.7.0 → 0.9.0

package/jobs/agriculture/equipment-maintenance-due-reminder.yaml

@@ -0,0 +1,36 @@
+id: equipment-maintenance-due-reminder
+version: 1
+name: Farm Equipment Maintenance Due Reminder
+description: >
+  Tracks engine hours or mileage against manufacturer service intervals for
+  tractors and other equipment, and warns when service is coming due. Use
+  this if equipment has ever broken down mid-season because a service
+  interval slipped by.
+category: agriculture
+tags: [equipment, maintenance, tractor, service]
+schedule: "0 6 * * 1"
+timezone: "UTC"
+runner: shell
+command: >
+  awk -F',' -v interval="{{service_interval_hours}}" -v warn="{{warn_hours_before}}" '
+    NR==1 {next}
+    {
+      hours_since_service = $2 - $3;
+      hours_until_due = interval - hours_since_service;
+      if (hours_until_due <= warn) {
+        printf "DUE SOON: %s — %d hrs since last service, %d hrs until next service\n", $1, hours_since_service, hours_until_due;
+      }
+    }
+  ' "{{equipment_hours_csv_path}}"
+variables:
+  equipment_hours_csv_path:
+    default: "./equipment-hours.csv"
+    description: "Path to a CSV with columns: equipment_name,current_hours,last_service_hours."
+  service_interval_hours:
+    default: 250
+    description: Manufacturer-recommended engine hours between services.
+  warn_hours_before:
+    default: 25
+    description: How many hours before the interval to start warning.
+compatible_agents: [generic, claude, codex]
+notes: Assumes a manually or telematics-updated CSV of current engine hours per machine — point it at whatever export your equipment's system provides.

package/jobs/agriculture/irrigation-schedule-check.yaml

@@ -0,0 +1,41 @@
+id: irrigation-schedule-check
+version: 1
+name: Irrigation Need Check
+description: >
+  Checks recent rainfall and forecast against a soil-moisture threshold to
+  tell you whether irrigation is actually needed today. Use this if you've
+  ever watered on a fixed schedule and either wasted water after rain or
+  under-watered during a dry stretch.
+category: agriculture
+tags: [irrigation, water, weather, crops]
+schedule: "0 6 * * *"
+timezone: "America/Chicago"
+runner: agent-prompt
+prompt: |
+  Check rainfall for the last {{lookback_days}} days and the forecast for
+  the next {{forecast_days}} days at latitude {{latitude}}, longitude
+  {{longitude}} (use any weather API available).
+  1. Sum recent rainfall (inches or mm) and note any significant rain
+     expected soon.
+  2. Compare against the crop's typical weekly water need of
+     {{weekly_need_inches}} inches.
+  3. Recommend: irrigate today, hold off (rain covers it), or irrigate
+     partially — with a one-line reason.
+variables:
+  latitude:
+    default: "40.0"
+    description: Latitude of the field.
+  longitude:
+    default: "-95.0"
+    description: Longitude of the field.
+  lookback_days:
+    default: 7
+    description: Days of rainfall history to sum.
+  forecast_days:
+    default: 3
+    description: Days of forecast to check before recommending irrigation.
+  weekly_need_inches:
+    default: 1.0
+    description: Approximate weekly water requirement for the crop in question.
+compatible_agents: [claude, codex, hermes, openclaw, generic]
+notes: A simple rainfall-vs-need heuristic — not a substitute for actual soil moisture sensors if you have them wired in.

package/jobs/agriculture/weather-frost-alert.yaml

@@ -0,0 +1,41 @@
+id: weather-frost-alert
+version: 1
+name: Frost/Freeze Risk Alert
+description: >
+  Checks the forecast for your fields and warns if frost or freeze
+  temperatures are coming, so you have time to cover crops or run
+  frost-protection irrigation. Use this if a surprise cold snap has ever
+  cost you a harvest.
+category: agriculture
+tags: [weather, frost, crops, alert]
+schedule: "0 18 * * *"
+timezone: "America/Chicago"
+runner: hybrid
+command: >
+  curl -s "https://api.open-meteo.com/v1/forecast?latitude={{latitude}}&longitude={{longitude}}&hourly=temperature_2m&forecast_days=2&temperature_unit=fahrenheit" |
+  grep -o '"temperature_2m":\[[^]]*\]' |
+  head -c 2000
+prompt: |
+  Fetch the next 48-hour hourly forecast for latitude {{latitude}}, longitude
+  {{longitude}} (use any weather API available).
+  1. Flag any hour where temperature is forecast at or below
+     {{frost_threshold_f}}°F.
+  2. Note the earliest and coldest such hour.
+  3. Give a plain recommendation: cover crops, run frost irrigation, or no
+     action needed, based on how low and how long the cold spell lasts.
+script_note: >
+  `command` dumps raw hourly temperature numbers with no threshold check or
+  recommendation — you'd still have to scan for the frost hours yourself.
+  `prompt` does the threshold comparison and gives an action recommendation.
+variables:
+  latitude:
+    default: "40.0"
+    description: Latitude of the field/farm location.
+  longitude:
+    default: "-95.0"
+    description: Longitude of the field/farm location.
+  frost_threshold_f:
+    default: 34
+    description: Temperature (°F) at or below which frost risk is a concern.
+compatible_agents: [claude, codex, hermes, openclaw, generic]
+notes: Uses the free Open-Meteo API in script mode (no key required); swap in your preferred ag-weather provider if you have one.

package/jobs/construction/change-order-approval-reminder.yaml

@@ -0,0 +1,31 @@
+id: change-order-approval-reminder
+version: 1
+name: Change Order Approval Reminder
+description: >
+  Flags change orders that have been waiting on client sign-off too long.
+  Use this if work has ever stalled, or gone ahead unapproved, because a
+  change order sat in someone's inbox for weeks.
+category: construction
+tags: [construction, change-orders, approvals]
+schedule: "0 8 * * 1-5"
+timezone: "UTC"
+runner: shell
+command: >
+  today=$(date +%s);
+  awk -F, -v today="$today" -v warn={{stale_days}} '
+    NR>1 && $3 == "pending" {
+      cmd="date -d \"" $2 "\" +%s 2>/dev/null || date -jf \"%Y-%m-%d\" \"" $2 "\" +%s";
+      cmd | getline sent_ts; close(cmd);
+      days_waiting = int((today - sent_ts) / 86400);
+      if (days_waiting >= warn) print "WARNING: change order \"" $1 "\" pending " days_waiting " days (sent " $2 ")";
+    }
+  ' {{change_orders_csv}}
+variables:
+  change_orders_csv:
+    default: "change-orders.csv"
+    description: "CSV with header row, columns: description,sent_date(YYYY-MM-DD),status(pending/approved/rejected)."
+  stale_days:
+    default: 5
+    description: Days pending before it's flagged for follow-up.
+compatible_agents: [generic, claude, codex]
+notes: Keep change-orders.csv updated as change orders go out and come back signed.

package/jobs/creator/crosspost-lag-check.yaml

@@ -0,0 +1,35 @@
+id: crosspost-lag-check
+version: 1
+name: Crosspost Lag Check
+description: >
+  Checks that content published on your primary platform actually got
+  mirrored to your other channels. Use this if a video or post has ever
+  gone up on YouTube and just never made it to the other platforms because
+  crossposting was manual and someone forgot.
+category: creator
+tags: [content, crosspost, publishing]
+schedule: "0 10 * * *"
+timezone: "UTC"
+runner: agent-prompt
+prompt: |
+  Compare recent uploads on {{primary_platform}} against what's been
+  posted on {{secondary_platforms}} over the last {{lookback_days}} days.
+  1. List content published on {{primary_platform}} with no matching post
+     on one or more of {{secondary_platforms}} within {{max_lag_days}}
+     days of the original.
+  2. Report each gap with the platform(s) missing it.
+variables:
+  primary_platform:
+    default: "YouTube"
+    description: Your main publishing platform.
+  secondary_platforms:
+    default: "Instagram, TikTok, X"
+    description: Comma-separated list of platforms content should also land on.
+  lookback_days:
+    default: 14
+    description: How far back to check for crossposting gaps.
+  max_lag_days:
+    default: 3
+    description: Days after the original post before a missing crosspost counts as a gap.
+compatible_agents: [claude, codex, hermes, openclaw, generic]
+notes: No generic script mode — checking multiple platform APIs and matching content across them needs an agent, not a fixed query.

package/jobs/crypto/gas-price-alert.yaml

@@ -0,0 +1,34 @@
+id: gas-price-alert
+version: 1
+name: Gas Price Alert
+description: >
+  Checks current network gas price and alerts when it drops below your
+  target, so you can time a planned transaction instead of overpaying
+  during a busy window. Use this if you've ever paid 5x normal gas
+  because you executed at a random moment instead of waiting.
+category: crypto
+tags: [crypto, gas, ethereum, blockchain]
+schedule: "*/15 * * * *"
+timezone: "UTC"
+runner: shell
+command: >
+  gas_hex=$(curl -s "{{rpc_endpoint}}" -X POST -H "Content-Type: application/json" \
+    --data '{"jsonrpc":"2.0","method":"eth_gasPrice","params":[],"id":1}' |
+    grep -o '"result":"0x[0-9a-fA-F]*"' | grep -o '0x[0-9a-fA-F]*');
+  if [ -n "$gas_hex" ]; then
+    wei=$((16#${gas_hex#0x}));
+    gwei=$(awk -v w="$wei" 'BEGIN{printf "%.2f", w/1000000000}');
+    echo "Current gas price: $gwei gwei";
+    awk -v g="$gwei" -v target="{{target_gwei}}" 'BEGIN{if (g+0 <= target+0) print "GAS BELOW TARGET — good time to transact"}';
+  else
+    echo "Could not fetch gas price — check RPC endpoint.";
+  fi
+variables:
+  rpc_endpoint:
+    default: "https://eth.llamarpc.com"
+    description: JSON-RPC endpoint for the chain to query.
+  target_gwei:
+    default: 20
+    description: Gas price (in gwei) at or below which to alert that it's a good time to transact.
+compatible_agents: [generic, claude, codex]
+notes: Read-only network query — no wallet/signing involved. Point at an L2's RPC endpoint if that's the chain you care about instead of mainnet.

package/jobs/crypto/smart-contract-upgrade-watch.yaml

@@ -0,0 +1,49 @@
+id: smart-contract-upgrade-watch
+version: 1
+name: Smart Contract Upgrade/Deprecation Watch
+description: >
+  Watches a tracked smart contract's on-chain events and any linked
+  project announcements for upgrades, migrations, or deprecation notices.
+  Use this if a protocol you depend on has ever migrated to a new
+  contract address and you found out only after something broke.
+category: crypto
+tags: [crypto, smart-contract, upgrade, monitoring]
+schedule: "0 */6 * * *"
+timezone: "UTC"
+runner: hybrid
+command: >
+  curl -s "{{rpc_endpoint}}" -X POST -H "Content-Type: application/json" \
+    --data "{\"jsonrpc\":\"2.0\",\"method\":\"eth_getCode\",\"params\":[\"{{contract_address}}\",\"latest\"],\"id\":1}" |
+    grep -o '"result":"0x[0-9a-fA-F]*"' | head -c 200
+prompt: |
+  Check contract {{contract_address}} on the relevant chain for recent
+  activity in the last {{lookback_hours}} hours (via a block explorer API
+  or RPC): new admin/upgrade events, proxy implementation changes, or
+  pause events.
+  Also check {{project_announcement_url}} for any migration or
+  deprecation announcement.
+  1. Report whether the contract's bytecode/implementation changed
+     recently (proxy upgrade) or an admin/ownership event fired.
+  2. Summarize any related announcement about migration, deprecation, or
+     a new contract address to switch to.
+  3. If nothing changed, say so in one line.
+script_note: >
+  `command` only confirms the contract still has code deployed at that
+  address — it can't detect proxy upgrades, admin events, or read
+  announcements. `prompt` checks actual upgrade/admin events and
+  cross-references project announcements.
+variables:
+  rpc_endpoint:
+    default: "https://eth.llamarpc.com"
+    description: JSON-RPC endpoint for the chain the contract lives on.
+  contract_address:
+    default: "0x0000000000000000000000000000000000000000"
+    description: Contract address to monitor.
+  project_announcement_url:
+    default: "https://example.com/blog"
+    description: Project blog/changelog/Twitter URL to check for migration announcements.
+  lookback_hours:
+    default: 6
+    description: How far back to check for on-chain events.
+compatible_agents: [claude, codex, hermes, openclaw, generic]
+notes: Read-only — no private key involved. Proxy-upgrade detection depends on the contract following a common proxy pattern (e.g. EIP-1967); adapt for non-standard contracts.

package/jobs/crypto/wallet-balance-watch.yaml

@@ -0,0 +1,37 @@
+id: wallet-balance-watch
+version: 1
+name: Wallet Balance Watch
+description: >
+  Checks a public wallet address's balance and alerts if it drops below a
+  threshold, without needing any private key or exchange API access. Use
+  this if you've ever wanted a heads-up when a hot wallet's gas/funding
+  balance ran low.
+category: crypto
+tags: [crypto, wallet, balance, blockchain]
+schedule: "*/30 * * * *"
+timezone: "UTC"
+runner: shell
+command: >
+  balance=$(curl -s "{{rpc_endpoint}}" -X POST -H "Content-Type: application/json" \
+    --data '{"jsonrpc":"2.0","method":"eth_getBalance","params":["{{wallet_address}}","latest"],"id":1}' |
+    grep -o '"result":"0x[0-9a-fA-F]*"' | grep -o '0x[0-9a-fA-F]*');
+  if [ -n "$balance" ]; then
+    wei=$((16#${balance#0x}));
+    eth=$(awk -v w="$wei" 'BEGIN{printf "%.6f", w/1000000000000000000}');
+    echo "Balance for {{wallet_address}}: $eth ETH";
+    awk -v bal="$eth" -v min="{{min_balance_eth}}" 'BEGIN{if (bal+0 < min+0) print "LOW BALANCE ALERT: below " min " ETH threshold"}';
+  else
+    echo "Could not fetch balance — check RPC endpoint/address.";
+  fi
+variables:
+  rpc_endpoint:
+    default: "https://eth.llamarpc.com"
+    description: JSON-RPC endpoint for the chain to query (public endpoint by default).
+  wallet_address:
+    default: "0x0000000000000000000000000000000000000000"
+    description: Public wallet address to monitor.
+  min_balance_eth:
+    default: 0.05
+    description: Balance threshold (in native token units) below which to alert.
+compatible_agents: [generic, claude, codex]
+notes: Read-only — public RPC calls only, no private key or signing involved. Swap the RPC endpoint and unit conversion for other EVM chains as needed.

package/jobs/devops/github-issue-triage.yaml

@@ -0,0 +1,38 @@
+id: github-issue-triage
+version: 1
+name: GitHub Issue Triage
+description: >
+  Reviews new GitHub issues opened since the last run and sorts them into
+  urgent/normal/low-priority with a suggested label and one-line reason.
+  Use this if issues pile up untriaged and the urgent ones get lost in the
+  noise until someone stumbles on them days later.
+category: devops
+tags: [github, issues, triage, devops]
+schedule: "0 8,14 * * 1-5"
+timezone: "UTC"
+runner: agent-prompt
+prompt: |
+  List issues opened in {{repo}} in the last {{lookback_hours}} hours (via
+  `gh issue list` or the GitHub API) that don't yet have a priority label.
+  1. For each, classify as urgent / normal / low-priority based on: does it
+     describe a crash, data loss, or security issue (urgent); a real bug
+     with a workaround (normal); or a feature request/question
+     (low-priority).
+  2. Give a one-line reason for the classification.
+  3. Suggest the matching label from: {{priority_labels}}.
+  4. List anything you'd flag for a human to look at today regardless of
+     label, separately at the top.
+  5. Do not apply labels or comment automatically — output a review-ready
+     list only.
+variables:
+  repo:
+    default: "owner/repo"
+    description: GitHub repo to triage, in owner/repo form.
+  lookback_hours:
+    default: 24
+    description: How far back to check for untriaged issues.
+  priority_labels:
+    default: "priority:urgent, priority:normal, priority:low"
+    description: Comma-separated label names to suggest, matching your repo's label scheme.
+compatible_agents: [claude, codex, hermes, openclaw, generic]
+notes: Read-only triage suggestions — wire up label/comment application separately once you trust the classifications, and requires `gh` CLI auth or API token with issue read access.

package/jobs/ecommerce/abandoned-cart-recovery-check.yaml

@@ -0,0 +1,38 @@
+id: abandoned-cart-recovery-check
+version: 1
+name: Abandoned Cart Recovery Check
+description: >
+  Reviews carts abandoned in the last day and drafts recovery emails for
+  ones above a minimum value, instead of relying only on your platform's
+  generic auto-recovery flow. Use this if you've ever suspected your
+  default abandoned-cart email is too generic to convert high-value carts.
+category: ecommerce
+tags: [ecommerce, cart-abandonment, recovery, sales]
+schedule: "0 10 * * *"
+timezone: "UTC"
+runner: agent-prompt
+prompt: |
+  Check {{store_platform}} for carts abandoned in the last
+  {{lookback_hours}} hours with a value at or above ${{min_cart_value}}.
+  1. List each cart: items, value, and how long ago it was abandoned.
+  2. Draft a personalized recovery email per cart referencing the actual
+     items (not a generic template), with a mild incentive if
+     {{allow_discount}} is true.
+  3. Flag any cart abandoned at the payment step specifically — these
+     often indicate a checkout friction problem worth investigating
+     separately from lack of interest.
+variables:
+  store_platform:
+    default: "your storefront platform"
+    description: Name of the e-commerce platform (e.g. "Shopify", "WooCommerce", "BigCommerce").
+  lookback_hours:
+    default: 24
+    description: How far back to check for abandoned carts.
+  min_cart_value:
+    default: 50
+    description: Minimum cart value worth a personalized recovery attempt.
+  allow_discount:
+    default: false
+    description: Whether the draft email may offer a discount incentive.
+compatible_agents: [claude, codex, hermes, openclaw, generic]
+notes: Draft-only — review before sending, and check your platform's marketing-consent rules before emailing anyone who didn't opt in.

package/jobs/ecommerce/low-stock-reorder-alert.yaml

@@ -0,0 +1,33 @@
+id: low-stock-reorder-alert
+version: 1
+name: Low Stock Reorder Alert
+description: >
+  Checks inventory levels against a reorder threshold per SKU and flags
+  anything that needs restocking before it actually sells out. Use this if
+  a bestseller has ever gone out of stock because nobody was watching the
+  count.
+category: ecommerce
+tags: [ecommerce, inventory, stock, reorder]
+schedule: "0 7 * * *"
+timezone: "UTC"
+runner: shell
+command: >
+  awk -F',' -v days="{{lead_time_days}}" '
+    NR==1 {next}
+    {
+      daily_velocity = $3 / 30;
+      days_of_stock = (daily_velocity > 0) ? $2 / daily_velocity : 9999;
+      if (days_of_stock <= days) {
+        printf "REORDER: %s — %d units left, ~%d days of stock at current velocity\n", $1, $2, days_of_stock;
+      }
+    }
+  ' "{{inventory_csv_path}}"
+variables:
+  inventory_csv_path:
+    default: "./inventory.csv"
+    description: "Path to a CSV with columns: sku,current_stock,units_sold_last_30_days."
+  lead_time_days:
+    default: 14
+    description: Supplier lead time in days — reorder when stock will run out before a new order could arrive.
+compatible_agents: [generic, claude, codex]
+notes: Velocity is a simple 30-day average — swap in a smarter forecast if your product has strong seasonality.

package/jobs/ecommerce/return-refund-backlog-check.yaml

@@ -0,0 +1,29 @@
+id: return-refund-backlog-check
+version: 1
+name: Return/Refund Backlog Check
+description: >
+  Checks how many return/refund requests are sitting unprocessed and flags
+  any past your stated turnaround time. Use this if refund complaints have
+  ever shown up in reviews before anyone internally noticed the backlog.
+category: ecommerce
+tags: [ecommerce, returns, refunds, backlog]
+schedule: "0 9 * * *"
+timezone: "UTC"
+runner: agent-prompt
+prompt: |
+  Check {{store_platform}} for return/refund requests still open.
+  1. Total open requests, grouped by how long they've been open.
+  2. Flag any request older than {{turnaround_sla_days}} days as past SLA.
+  3. For past-SLA requests, draft a short customer-facing update
+     apologizing for the delay and giving a status/next step.
+  4. Summarize the most common return reason across the backlog, in case
+     it points to a product quality issue worth escalating.
+variables:
+  store_platform:
+    default: "your storefront/returns platform"
+    description: Name of the platform to check (e.g. "Shopify", "Loop Returns", "Happy Returns").
+  turnaround_sla_days:
+    default: 3
+    description: Your stated turnaround time for processing a return/refund.
+compatible_agents: [claude, codex, hermes, openclaw, generic]
+notes: Draft-only — review customer-facing messages before sending, especially any that touch on refund amounts or timing promises.

package/jobs/education/report-card-deadline-reminder.yaml

@@ -0,0 +1,27 @@
+id: report-card-deadline-reminder
+version: 1
+name: Report Card Deadline Reminder
+description: >
+  Warns you before the grade submission deadline for report cards. Use
+  this if grades have ever gone in late because the cutoff snuck up
+  during the last busy week of the term.
+category: education
+tags: [education, grading, deadlines]
+schedule: "0 8 * * *"
+timezone: "UTC"
+runner: shell
+command: >
+  today=$(date +%s);
+  due_ts=$(date -d "{{grades_due_date}}" +%s 2>/dev/null || date -jf "%Y-%m-%d" "{{grades_due_date}}" +%s);
+  days_left=$(( (due_ts - today) / 86400 ));
+  if [ "$days_left" -le "{{warn_days}}" ] && [ "$days_left" -ge 0 ]; then echo "WARNING: grades due in $days_left days ({{grades_due_date}})"; exit 1;
+  else echo "OK: grades due in $days_left days"; fi
+variables:
+  grades_due_date:
+    default: "2026-12-15"
+    description: Grade submission deadline for this term (YYYY-MM-DD).
+  warn_days:
+    default: 5
+    description: Days before the deadline to start warning.
+compatible_agents: [generic, claude, codex]
+notes: Manual grades_due_date input — update it each term based on the school's academic calendar.

package/jobs/events/event-budget-overrun-check.yaml

@@ -0,0 +1,31 @@
+id: event-budget-overrun-check
+version: 1
+name: Event Budget Overrun Check
+description: >
+  Compares committed spend so far against the event budget and flags
+  overrun risk. Use this if an event has ever come in over budget because
+  nobody totaled up the vendor contracts until after the fact.
+category: events
+tags: [events, budget, planning]
+schedule: "0 9 * * 1"
+timezone: "UTC"
+runner: shell
+command: >
+  spent=$(awk -F, 'NR>1 {sum+=$2} END {print sum+0}' {{expenses_csv}});
+  budget={{budget_usd}};
+  if awk "BEGIN{exit !($spent >= $budget * {{warn_pct}} / 100)}"; then
+    echo "WARNING: committed spend \$$spent has hit ${warn_pct:-{{warn_pct}}}% of \$$budget budget";
+    exit 1;
+  else echo "OK: committed spend \$$spent of \$$budget budget"; fi
+variables:
+  expenses_csv:
+    default: "event-expenses.csv"
+    description: "CSV with header row, columns: line_item,amount."
+  budget_usd:
+    default: 10000
+    description: Total event budget in USD.
+  warn_pct:
+    default: 90
+    description: Percent of budget committed before it's flagged.
+compatible_agents: [generic, claude, codex]
+notes: Keep event-expenses.csv updated as vendor contracts and deposits are committed, not just when invoices arrive.

package/jobs/finance/subscription-price-hike-watch.yaml

@@ -0,0 +1,44 @@
+id: subscription-price-hike-watch
+version: 1
+name: Subscription Price Hike Watch
+description: >
+  Checks a tracked subscription's current price against the last price you
+  recorded and flags any increase. Use this if a service has ever quietly
+  raised its price and you only noticed months later scrolling through a
+  bank statement.
+category: finance
+tags: [subscriptions, price, billing, finance]
+schedule: "0 8 1 * *"
+timezone: "UTC"
+runner: shell
+command: >
+  history_file="{{price_history_path}}";
+  touch "$history_file";
+  current_price=$(curl -s -A "Mozilla/5.0" "{{pricing_page_url}}" | grep -o '{{price_pattern}}' | head -1);
+  if [ -z "$current_price" ]; then
+    echo "could not extract a price from {{pricing_page_url}} — check {{price_pattern}} still matches the page.";
+    exit 0;
+  fi
+  last_price=$(tail -1 "$history_file" | cut -d',' -f2);
+  today=$(date +%Y-%m-%d);
+  echo "$today,$current_price" >> "$history_file";
+  if [ -n "$last_price" ] && [ "$current_price" != "$last_price" ]; then
+    echo "PRICE CHANGED for {{service_name}}: $last_price -> $current_price";
+  else
+    echo "{{service_name}} price unchanged: $current_price";
+  fi
+variables:
+  service_name:
+    default: "your subscription service"
+    description: Name of the subscription being tracked.
+  pricing_page_url:
+    default: "https://example.com/pricing"
+    description: Public pricing page URL to check.
+  price_pattern:
+    default: "\\$[0-9]+\\.[0-9][0-9]"
+    description: A grep pattern that extracts the price text from the page (adjust per site's markup).
+  price_history_path:
+    default: "./subscription-price-history.csv"
+    description: Local CSV file to append price observations to over time.
+compatible_agents: [generic, claude, codex]
+notes: Extraction is a simple grep on rendered HTML — pages that render price via JS won't match; use an agent-prompt job or a headless browser if that's the case for your target service.

package/jobs/fleet/vehicle-registration-expiry-check.yaml

@@ -0,0 +1,31 @@
+id: vehicle-registration-expiry-check
+version: 1
+name: Fleet Vehicle Registration Expiry Check
+description: >
+  Warns you before a fleet vehicle's registration expires. Use this if a
+  vehicle has ever gotten pulled over, or failed a job site check-in,
+  because its registration quietly lapsed.
+category: fleet
+tags: [fleet, registration, compliance]
+schedule: "0 8 1 * *"
+timezone: "UTC"
+runner: shell
+command: >
+  today=$(date +%s);
+  awk -F, -v today="$today" -v warn={{warn_days}} '
+    NR>1 {
+      cmd="date -d \"" $2 "\" +%s 2>/dev/null || date -jf \"%Y-%m-%d\" \"" $2 "\" +%s";
+      cmd | getline exp_ts; close(cmd);
+      days_left = int((exp_ts - today) / 86400);
+      if (days_left <= warn) print "WARNING: vehicle " $1 " registration expires in " days_left " days (" $2 ")";
+    }
+  ' {{fleet_registration_csv}}
+variables:
+  fleet_registration_csv:
+    default: "fleet-registration.csv"
+    description: "CSV with header row, columns: vehicle_id,registration_expiry(YYYY-MM-DD)."
+  warn_days:
+    default: 30
+    description: Days before expiry to start warning.
+compatible_agents: [generic, claude, codex]
+notes: Keep fleet-registration.csv updated after each renewal — one row per vehicle.

package/jobs/gaming/game-patch-notes-watch.yaml

@@ -0,0 +1,39 @@
+id: game-patch-notes-watch
+version: 1
+name: Game Patch Notes Watch
+description: >
+  Checks whether any game you regularly stream or play competitively has
+  shipped a new patch, and summarizes what changed. Use this if you've
+  ever gone live only to discover mid-stream that a balance patch dropped
+  and your strategies are stale.
+category: gaming
+tags: [gaming, patch-notes, updates, competitive]
+schedule: "0 8 * * *"
+timezone: "UTC"
+runner: hybrid
+command: >
+  for url in {{patch_notes_urls}}; do
+    echo "--- $url ---";
+    curl -s -A "Mozilla/5.0" "$url" | grep -io '<title>[^<]*</title>' | head -1;
+  done
+prompt: |
+  Check each of these patch-notes sources for updates published in the
+  last {{lookback_hours}} hours: {{patch_notes_urls}}.
+  1. For any source with a new patch, summarize the top 3-5 changes most
+     relevant to competitive play or content strategy (balance changes,
+     new content, meta-shifting changes) — skip pure bugfix trivia.
+  2. Note the patch version/date.
+  3. If nothing new, say so in one line per source.
+script_note: >
+  `command` just fetches each page's title to spot an update, with no
+  actual summary of what changed — you'd still have to read the patch
+  notes yourself. `prompt` reads and summarizes the meaningful changes.
+variables:
+  patch_notes_urls:
+    default: "https://www.leagueoflegends.com/en-us/news/tags/patch-notes/"
+    description: Space-separated list of patch-notes page URLs to watch.
+  lookback_hours:
+    default: 24
+    description: How recent a patch needs to be to count as "new."
+compatible_agents: [claude, codex, hermes, openclaw, generic]
+notes: Works best for games with a stable patch-notes URL/RSS feed; some publishers gate notes behind launcher clients instead of a webpage.

package/jobs/gaming/stream-schedule-reminder.yaml

@@ -0,0 +1,30 @@
+id: stream-schedule-reminder
+version: 1
+name: Stream Schedule Prep Reminder
+description: >
+  Reminds you before a scheduled stream and checks that your streaming
+  software/overlays are up to date. Use this if you've ever gone live late
+  because you were still fighting an OBS update mid-countdown.
+category: gaming
+tags: [streaming, schedule, obs, reminder]
+schedule: "0 * * * *"
+timezone: "America/Los_Angeles"
+runner: shell
+command: >
+  now=$(date +%H:%M);
+  target="{{stream_time}}";
+  echo "Now: $now, next stream: $target on {{stream_days}}.";
+  if command -v obs >/dev/null 2>&1; then
+    obs --version 2>/dev/null || echo "OBS installed but version check failed";
+  else
+    echo "OBS CLI not found on PATH — check your streaming software manually.";
+  fi
+variables:
+  stream_time:
+    default: "18:00"
+    description: Scheduled stream start time (24h, local timezone).
+  stream_days:
+    default: "Mon/Wed/Fri"
+    description: Days you typically stream, for the reminder text.
+compatible_agents: [generic, claude, codex]
+notes: Runs hourly and just prints the target time for a human/agent to compare — wire it to your actual OBS/streaming software's API for a real go/no-go check if available.