Skip to content

ZenUml/conf-app / .claude/skills/income-radar

income-radar

Verified

Near-term Atlassian Marketplace cash view for the ZenUML apps — what vendor income to expect over the next few days (client subscriptions renewing) and which paying clients missed a payment in the past few days (renewal lapsed / in grace / dead payment method). Use when the user asks "what income am I expecting", "what's landing this week", "any renewals coming up", "did anyone miss a payment", "who's late", "income projection for the next few days", or wants a daily payments radar. Built on the marketplace engine (licenses + transactions joined on cloudId), payers only, all revenue apps. For lifetime revenue / top payers / a full renewals-in-a-window audit use `marketplace`; for a single tenant's paid status use `tenant`.

Income Radar

A daily near-term cash view built on top of the marketplace engine: what's about to land, and what already slipped.

Quick start

# from the conf-app repo root (creds auto-load from .env.forge.local)
S=.claude/skills/marketplace/scripts/mp_report.py

python3 $S radar                 # default: +/-7 days around today
python3 $S radar --days 14       # wider window
python3 $S radar --asof 2026-07-01 --json
python3 $S radar --local         # offline, from the last `sync` snapshot (fast, no creds)

--local needs a snapshot first (python3 $S sync, ~14s, run from marketplace).

What it reports

Four sections:

  • INCOMING — renewals due in the next N days, summed as expected income. Each row carries the same payment-risk flags as MISSED, and the header splits out how much of the expected total sits on a flagged/dead card (a renewal on no-payment-method is the leading indicator it will fail and land in MISSED next cycle).
  • MISSED — payer renewals that lapsed in the past N days without a new payment landing, summed as income at risk. A license still in its grace period is excluded (see IN GRACE).
  • IN GRACE — paid coverage ended (paid_thru < today) but the license is active with inGracePeriod = Yes and its maintenanceEndDate (the grace deadline, grace_until) is today or later. The renewal is pending, not missed. Listed regardless of the look-back window until the grace deadline passes; its total is separate from MISSED.
  • EVALUATIONS — trials (licenseType = EVALUATION) expiring in the next N days (a billing observation window — not an instruction to contact them) or that expired in the past N days. A lifecycle signal, not income — shown in its own section, kept out of the $ totals. Each row shows converted (does this tenant+app already have a paid transaction). Lite is excluded (free, no paid conversion).

Scope: all revenue apps (Full + Diagramly + AsyncAPI; Lite is a free Marketplace listing, so it's ~$0 here — its paid layer is Stripe/KV, see extend-space-license / paywall). Payers only — clients with lifetime vendor $ > 0. A never-paying trial or free install never shows up in either bucket.

Before presenting these candidates as a paid cohort, verify explicit source-reported Paid/Fully paid status in raw transactions: the current helper's amount-based summaries do not perform that check. Open-only orders are purchase evidence with unconfirmed settlement, not received income. A historical payer's newer Open renewal is separate from an absent renewal; do not call that customer “not purchased” or automatically chase them. Overlapping invoices or unresolved refunds can make totals uncertain.

One row per (tenant, app) subscription — a tenant that pays for both Full and Diagramly gets two rows (its two real renewals), but a tenant with a paid Full license plus a leftover free Lite listing gets one (the empty Lite bucket is $0 and drops out — this is why revenue is aggregated by (cloudId, addonKey), not cloudId alone, so a phantom Lite row can't double-count the Full income). ZenUML's own internal instances (zenuml, zenuml-connect) are excluded.

Definitions

  • paid_thru — the helper's latest positive-amount transaction end; verify raw status before calling it settled coverage. A future start is not coverage today. This is not a vendor payout date.
  • expected amount — the vendorAmount of that client's most recent paid transaction (the number most likely to repeat on the next cycle).
  • INCOMING: today <= paid_thru <= today + N.
  • IN GRACE: paid_thru < today, inGracePeriod = Yes, today <= grace_until, where grace_until is the license maintenanceEndDate. Checked before INCOMING and MISSED.
  • MISSED: today - N <= paid_thru < today and not IN GRACE — no newer paid transaction has extended coverage past today. Reinforced with invoiceDunningReason ("no-payment-method") flags, plus inactive if the license itself went inactive.

Output

Default (text): a header, then five sections, each a headline + a padded table. Empty sections print (none). In --local mode a [local snapshot @ … (Nh old)] line goes to stderr, keeping stdout clean.

=== income radar  asof=YYYY-MM-DD  window=+/-Nd  (all revenue apps, payers only) ===

INCOMING (next Nd): C renewals, ~ $T expected  (at-risk on flagged cards: $R)
  due  app  billing  amount  flags  tier  company  host  entitlement
MISSED (past Nd): C renewals overdue, ~ $T at risk
  paid_thru  days_late  app  billing  amount  flags  company  host  entitlement
IN GRACE: C renewals pending (paid coverage ended, license still active), ~ $T
  grace_until  days_left  paid_thru  app  billing  amount  flags  company  host  entitlement
note: … (3 lines on renewal-timing-as-proxy)
EVALUATIONS expiring (next Nd): C trials — conversion window (not income)
  expires  app  tier  status  converted  company  host  entitlement
EVALUATIONS expired (past Nd): C trials
  expires  days_ago  app  tier  converted  company  host  entitlement

--json: one object. Dollar totals live only in the income sections; evaluations carry no total by design (trials are $0). Each rows[] entry has the fields of its section's columns.

{
  "asof": "YYYY-MM-DD", "days": N,
  "incoming": { "total": <$>, "at_risk": <$>, "count": <n>, "rows": [ { "company","host","entitlement","app","billing","amount","tier","flags","due" } ] },
  "missed":   { "total": <$>, "count": <n>, "rows": [ { …, "paid_thru", "days_late" } ] },
  "in_grace": { "total": <$>, "count": <n>, "rows": [ { …, "paid_thru", "grace_until", "days_left" } ] },
  "evaluations": {
    "expiring": { "count": <n>, "rows": [ { "expires","app","tier","status","converted","company","host","entitlement" } ] },
    "expired":  { "count": <n>, "rows": [ { …, "days_ago" } ] }
  }
}

Caveats

Dates are customer-renewal timing — a proxy for income, not a disbursement schedule. Atlassian pays out vendors on its own monthly cycle, independent of any one client's renewal date. A lapse in the last 1-2 days may just be settlement lag, not a true miss — read the grace / no-payment-method flags before treating a fresh MISSED row as lost revenue.

Fast path (--local)

--local reads the shared marketplace SQLite snapshot instead of hitting the live API. Run python3 $S sync first to build/refresh it. The command prints the snapshot age on stderr and warns past 24h — for a "what does today actually look like" answer, run live instead (radar is billing-sensitive, unlike the cloudId↔domain identity that snapshot lookups elsewhere rely on).

Related

  • marketplace — the shared engine (mp_report.py, sync), lifetime revenue, top payers, and a full renewals-in-a-window audit (renewals --from --to).
  • tenant — single-tenant lookup ("is <domain> paying, how big, what state").
  • extend-space-license / paywall — the Lite Stripe/KV layer; Lite paid access isn't in Marketplace transactions, so it never appears in this radar.

Follow-up boundary

Read-only facts and gaps. A trial deadline alone is not urgency, a failed automatic payment or authority to contact. Use customer-lifecycle for state changes and customer-followup for decisions. Marketplace all may include vendor products beyond this document; report the actual product scope.

rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall