Read-only repo triage · 13 Oct 2026 local snapshot

Your repos: mess vs money

I checked the workspace as a portfolio, then read the strongest-looking sales candidates more closely. The short answer: your fastest plausible money is a service sold through Igloo Interactive—not another SaaS launch.

Start with igloointeractive.ca. The site already presents defined, priced work and a way to contact you. One $50–$75/hour technical review or a small $300 website job beats waiting for strangers to discover a new app. That is one-off revenue, not recurring monthly revenue; the repo contains no proof of inbound leads or customers.

If you mean recurring CAD $100/month specifically, NightDesk is the closest fit—but it is not at paid-customer readiness yet. Validate and hand-onboard one real business before adding billing or more features.

Best overlap of “less mess” and “money soon”

1igloointeractive.ca

Clean-ish checkoutSell services now

The strongest combined bet. Astro marketing site, a clear three-part offer, published entry prices ($300 web/setup; $50–$75/hour reviews), a working-looking enquiry path, and a clean Git tree. No payments or SaaS machinery are needed to invoice for a service.

What stands in the way: customer acquisition and proof, not another round of site polishing. The project handoff notes no attributable testimonial; analytics/Search Console decisions are still open. I found no lead or revenue data.

Smallest test: sell one tightly scoped paid review or starter site to an actual local business this week. Ask what they already pay to solve the problem; don't count compliments or waitlist signups as validation.

2nightdesk.ca

Best subscription upsideNot payment-ready

Real call-screening, voicemail transcription, SMS, escalation and Twilio/OpenRouter integration—not just a landing page. Small repo and clean tree. Of the software candidates this is the clearest recurring B2B offer.

Why it is not first: there are no automated tests in the checkout; user accounts and Stripe billing are still unchecked milestones. The roadmap says the custom domain is not connected, and its handoff calls for a real end-to-end call test. Site pricing promises a lot, but those capabilities are not evidence of paid, reliable service.

Smallest test: pre-sell a supervised, manually onboarded after-hours call-handling pilot to one contractor. Verify real call routing, escalation and failure handling with that customer before promising unattended service. Don't build billing first.

3refcheck.cv

Clear transactional priceCheckout closed

Compact, coherent marketing product, sample report, legal pages and a clean tree; listed packages run from $19.99 to $99.99. But the public purchase path is deliberately disabled: checkout is a static notice, Stripe/order creation are unfinished, and the notice still says to return at the end of August. It is now October in the local project records.

Blockers: no active order/fulfilment flow or billing, no tests in the repo, and candidate reference checks have meaningful consent, privacy and reputation risk. Get qualified legal review of the actual process and claims before accepting paid orders.

Smallest test: update the stale availability message, then validate demand with a real, consented, manual pilot and a lawful, reviewed script. Don't sell the paused checkout as live.

4callverify.ca

Payments exist in codeStale / operationally heavy

The most tangible payment implementation among the small-business ideas: Laravel login, a credit wallet, call workflows and Stripe checkout/webhook code are present. That makes it more than a mockup.

Why it falls down: last application commit is May 2026; tests look like framework examples, deployment requires several services, and the recording/consent/compliance story raises the cost of a safe launch. A code path for Stripe is not proof a production payment and call have succeeded.

Before choosing it: get one exact local/production build and call/payment walkthrough working, then review recording disclosures and legal boundaries. Otherwise it is a larger rescue project than a fast $100 test.

The rest of the money shortlist

Repo / bet	Relative mess	Fast-money verdict
lastwords Digital dead-man's switch	Medium: real app code, but its own plan still lists session-cookie, scheduled delivery, Stripe, launch, privacy/legal, security and recovery work.	Not quick. High trust burden: a false or missed message delivery can seriously hurt someone. Do not market the reliability/zero-knowledge/non-profit guarantees before they are implemented and independently checked.
affy Affiliate-site portfolio	Large content surface; repository plans flag affiliate validation, placeholder links and launch blockers. README's six-blog description and plan's nine-blog portfolio disagree.	Potential longer-term experiment, not near-term cash. Partner approval, genuine links, search distribution and conversion evidence are missing. More articles alone won't make it earn.
igloonetworks.com/ivr Twilio workflow control plane	Large/complex, but notably disciplined: substantial automated tests and explicit safety gates. Current checkout is on a feature branch; live Twilio staging remains unproven in the plan.	Strong engineering work, weak immediate monetization. Admin control plane, local SQLite assumption, staging and customer deployment still stand between code and a self-serve product.
dashdia Local ideas dashboard	Relatively focused and clean, with tests/typecheck/build commands documented.	Good internal tool, not yet an external product: local single-user scope, no auth/tenant boundary or hosted offer. Don't confuse personal usefulness with willingness to pay.
snipurl.cc / vidput.com URL shortener / video hosting	Small MVP / older Laravel app respectively; both date from January 2026.	Park them. Shorteners compete with free incumbents; video storage, bandwidth, processing and abuse moderation create costs before revenue.

So what should you do?

Next 7 days: get paid for a service

Pick one audience and one plain offer from the existing Igloo site.
Contact 10 businesses you can actually reach; offer a paid, fixed-scope review or setup—not free custom development.
Ask for a deposit before custom work. Record every reply, objection and paid conversion in one simple ledger.
Only change the site after real prospects show you what is confusing.

If MRR matters: test NightDesk by hand

Find one contractor with a real missed-call/after-hours problem.
Agree on a limited pilot, price, call path, escalation fallback and who responds when AI fails.
Run real test calls and inspect logs; stay present and keep a human fallback.
Ask for payment after the pilot only if it solves a recurring problem. Automate onboarding/billing after that evidence.

The uncomfortable bit: your bottleneck is not a shortage of repositories. It is distribution and proof that a buyer will pay. Don't spend the next month making repo #55 while the only repository with a clear paid service funnel waits for prospects.

Portfolio mess, in plain English

I found 54 Git checkout directories in the workspace tree (including nested/experimental checkouts). That is not 54 independent businesses. There are symlink aliases, duplicated/personal archives, sysadmin dashboards, notes repositories, prototypes and unfinished products.

Best business surface: igloointeractive.ca; use this as the shopfront rather than launching another domain/product.
One product worth validating rather than broadening: nightdesk.ca, as a supervised service before a self-serve SaaS.
Worktree handoff friction: dasher had 87 changed/untracked paths, models 126, justthetips 15, talanovdb 15, and askai-creative 7 in this snapshot. Some may be intentional work; don't discard or clean them blindly.
Docs drift: several sales candidates preserve launch dates/pricing/status from older plans. Treat checked boxes and revenue forecasts as claims to verify, not market proof.
Evidence missing across this review: no repository I inspected contained reliable current paid-user, sales-conversion or revenue numbers. That doesn't prove there are no external customers; it means I can't rank on traction.

How I judged it

I weighted things a buyer can act on now: a coherent offer, current contact or checkout path, how much code/operations remain before safe fulfilment, local Git/worktree state, and whether the repo contains tests or current operational evidence. I gave zero credit for market-size estimates, predicted conversion rates, stale roadmaps, or “already live” claims without evidence. The scores are comparative judgement, not forecasts.

Scope: read-only inspection of repository inventory/status, manifests, READMEs, project plans/handoffs, and relevant sales/checkout/code paths. I did not install dependencies, run builds/tests, contact customers, verify public production endpoints, or inspect payment/usage analytics. Several repositories are still “metadata-triaged,” not full line-by-line code reviewed.