OPS ► REVENUE & COSTS
Load ops-rules before acting. Public repo (no personal data). Outbound: one draft → one approval → one send. If AskUserQuestion / Workflow are missing, follow Rule 10 in ops-rules (Hermes: numbered options / two-turn Telegram card; delegate_task).
Runtime Context
Before executing, load available context:
-
Preferences: Read
${CLAUDE_PLUGIN_DATA_DIR:-$HOME/.claude/plugins/data/ops-ops-marketplace}/preferences.jsontimezone— display all timestamps correctly
-
Daemon health: Read
${CLAUDE_PLUGIN_DATA_DIR}/daemon-health.json- If
action_neededis not null → surface it before the cost report
- If
-
Secrets: AWS Cost Explorer requires credentials.
Secret Resolution
- AWS: check
$AWS_PROFILE/$AWS_ACCESS_KEY_ID→doppler secrets get AWS_ACCESS_KEY_ID --plain→ vault query cmd from prefs - If no credentials available, report "AWS costs unavailable — credentials not configured" and show only the revenue pipeline from registry
- AWS: check
CLI/API Reference
AWS burn doctrine (CRITICAL — 2026-07-22)
Never report plain UnblendedCost / NetUnblendedCost totals as "spend".
Startup credits, SPP discounts, and Savings Plan negation make those nets ≈ $0
while real Usage burn is hundreds/day. Always use RECORD_TYPE=Usage for
burn, pace, spikes, runway, and Bedrock. Show credits only as a mask line.
Canonical helper (prefer this over raw aws ce):
${CLAUDE_PLUGIN_ROOT}/scripts/aws-usage-cost.sh snapshot 2>/dev/null || echo '{}'
${CLAUDE_PLUGIN_ROOT}/scripts/aws-usage-cost.sh all 2>/dev/null || true
| Command | Usage | Output |
|---|---|---|
scripts/aws-usage-cost.sh snapshot | MTD/prev Usage burn + 7d daily + top services + credit mask | JSON |
scripts/aws-usage-cost.sh daily 7 | Last N days Usage totals | JSON array |
scripts/aws-usage-cost.sh by-service mtd | MTD Usage by service | JSON array |
scripts/aws-usage-cost.sh record-types | Credit / Usage / SP breakdown (mask only) | JSON array |
aws ce get-cost-and-usage … --filter '{"Dimensions":{"Key":"RECORD_TYPE","Values":["Usage"]}}' | Raw Usage-only CE (fallback if helper missing) | Cost JSON |
aws ce list-savings-plans-purchase-recommendation | Savings plan recommendations | JSON |
Phase 1 — Gather financial data in parallel
FinOps dashboard — canonical revenue + burn
If FINOPS_DASHBOARD_URL and FINOPS_OPS_API_TOKEN are configured, the
dashboard is the single source of truth for per-project MRR, 30d revenue,
and current burn. Per-DBA breakdown comes from the Stripe revenue
snapshots ingested via /api/ops/revenue. Falls open to {} if unset —
the per-source AWS / RevenueCat queries below run as fallback.
If the dashboard current_month_spend is ≈ $0 while Usage burn from
aws-usage-cost.sh is material, trust the helper (dashboard may still be
credit-masked or ingest-empty). Always dual-check with the Usage helper.
${CLAUDE_PLUGIN_ROOT}/scripts/finops-bridge.sh snapshot 2>/dev/null || echo "{}"
${CLAUDE_PLUGIN_ROOT}/scripts/finops-bridge.sh revenue project 2>/dev/null || echo "{}"
When the dashboard returns data, render the per-project MRR section from
groups[] rather than from RevenueCat-only data.
AWS costs — Usage burn (authoritative)
${CLAUDE_PLUGIN_ROOT}/scripts/aws-usage-cost.sh snapshot 2>/dev/null || echo '{}'
Fallback if helper missing:
bashUSAGE_FILTER='{"Dimensions":{"Key":"RECORD_TYPE","Values":["Usage"]}}' aws ce get-cost-and-usage \ --time-period "Start=$(date +%Y-%m-01),End=$(date +%Y-%m-%d)" \ --granularity MONTHLY \ --metrics "UnblendedCost" \ --filter "$USAGE_FILTER" \ --group-by "Type=DIMENSION,Key=SERVICE" \ --output json 2>/dev/null
AWS costs (last 3 months Usage trend)
bashUSAGE_FILTER='{"Dimensions":{"Key":"RECORD_TYPE","Values":["Usage"]}}' aws ce get-cost-and-usage \ --time-period "Start=$(date -v-3m +%Y-%m-01 2>/dev/null || date -d '3 months ago' +%Y-%m-01),End=$(date +%Y-%m-%d)" \ --granularity MONTHLY \ --metrics "UnblendedCost" \ --filter "$USAGE_FILTER" \ --output json 2>/dev/null
AWS credits / discount mask (not spend)
${CLAUDE_PLUGIN_ROOT}/scripts/aws-usage-cost.sh record-types 2>/dev/null || echo '[]'
bashaws ce list-savings-plans-purchase-recommendation --output json 2>/dev/null || echo '{}' aws ce get-credits --output json 2>/dev/null || echo "credits API unavailable"
AWS cost forecast (end of month)
Note: CE forecasts are often net-of-credits. Prefer eom_pace from
aws-usage-cost.sh snapshot (7d Usage avg × days in month) as the burn forecast.
bashaws ce get-cost-forecast \ --time-period "Start=$(date +%Y-%m-%d),End=$(date +%Y-%m-28)" \ --metric "UNBLENDED_COST" \ --granularity MONTHLY \ --output json 2>/dev/null
Project registry (revenue stage)
bashcat "${CLAUDE_PLUGIN_ROOT}/scripts/registry.json" 2>/dev/null | jq '[.projects[] | {alias, name, stage: (.revenue_stage // .revenue.stage // "pre-revenue"), mrr: (.mrr // 0), source: (.source // "git"), type: (.type // "repo")}]'
External project revenue (Shopify, custom SaaS)
${CLAUDE_PLUGIN_ROOT}/bin/ops-external 2>/dev/null || echo '[]'
For Shopify projects showing status: "healthy", pull GMV via Shopify Admin API:
bash# For each Shopify project in registry with valid credentials: STORE_URL="[from project.shopify.store_url]" TOKEN="[from env var named in project.shopify.credential_key]" curl -s -H "X-Shopify-Access-Token: $TOKEN" \ "https://$STORE_URL/admin/api/2024-10/orders.json?status=any&created_at_min=$(date -v-30d +%Y-%m-%dT00:00:00Z 2>/dev/null)&limit=250" 2>/dev/null
Include Shopify GMV in the revenue pipeline table with source=shopify.
Phase 2 — Render dashboard
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ OPS ► REVENUE & COSTS — [month] ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ AWS USAGE BURN (RECORD_TYPE=Usage — not credit-masked net) This month usage: $[X] ← from aws-usage-cost.sh mtd_usage Forecast (EOM pace): $[X] ← eom_pace (7d Usage avg × days) Last month usage: $[X] MoM change: [+/-X%] 7d avg/day: $[X] Top services (Usage): [service] $[X] ([%] of usage) [service] $[X] ... CREDITS / MASK (do not report as low spend) Credits applied MTD: $[X] (record_types Credit) Discounts MTD: $[X] Net after mask: ~$0 is EXPECTED when credits cover burn Credits remaining: $[X] (if API available) Expires: [date] Months cover at pace: [N] REVENUE PIPELINE PROJECT SOURCE STAGE MRR/GMV STATUS ────────────────────────────────────────────────────────────── [project] git [stage] $[X] [status] [project] shopify [stage] $[X] GMV [status] [project] custom [stage] $[X] [status] ... ────────────────────────────────────────────────────────────── TOTAL MRR $[X] TOTAL SHOPIFY GMV (30d) $[X] RUNWAY ESTIMATE Monthly burn (AWS): $[X] Total MRR: $[X] Net burn: $[X/month] Credits cover: [N months] Cash runway: [depends on external data] ──────────────────────────────────────────────────────
Use batched AskUserQuestion calls (max 4 options each):
AskUserQuestion call 1:
[Drill into AWS costs by service] [Show cost anomalies (spike detection)] [Export cost report] [More...]
AskUserQuestion call 2 (only if "More..."):
[Update project revenue stage] [Set budget alert]
Route by $ARGUMENTS
| Argument | Action |
|---|---|
| costs | Show only AWS cost breakdown |
| credits | Show only credits and expiry |
| revenue | Show only revenue pipeline |
| runway | Calculate and show runway |
| (empty) | Show full dashboard |
Use AskUserQuestion after the dashboard for next action.
Native tool usage
WebFetch — billing API fallback
When aws ce commands fail or return incomplete data, use WebFetch to query the AWS Cost Explorer API directly. Also useful for fetching Stripe/billing provider data if configured.
Write — export reports
When user selects "Export cost report" (option c), use Write to save the report as a dated file:
Write(file_path: "/tmp/ops-revenue-[date].md", content: "[formatted report]")
