Workflows, agents, integrations — n8n, OpenAI and Anthropic APIs, your own stack — delivered with the same security review baked in. The part that reads your inbox and moves your money is the part that gets attacked first.
from ~$1,200 · written scope · fixed price · on your accounts
The boring job, done by a machine that can't be talked into things.
Built on the tools on the home page: n8n, the OpenAI and Anthropic APIs, Supabase, Postgres, Vercel, AWS — in your accounts, in your name.
01
Workflow automations
Intake, triage, follow-ups, reports. An event comes in, a model reads it, the right thing happens, a person is asked only when it matters.
02
Agents with tools
A model that can look things up and take actions — against your CRM, your database, your calendar — with least privilege and a confirmation step on anything that spends or deletes.
03
Integrations
The glue between the tools you already pay for: webhooks, APIs, queues, the odd scraper. Credentials stored properly, retries and rate limits handled.
04
Internal tools with a model in them
A small app around the workflow: a queue, an approval screen, a search over your documents. Next.js and Supabase, like the builds.
05
Moving an existing automation onto your accounts
Off a freelancer's n8n instance or a personal API key and into infrastructure you own, with the keys rotated on the way.
how it goes
Scope, build, attack, hand over.
The same engagement as everything else on this site — the only thing that changes is what I'm attacking at the end.
01
A free 30-minute call, then a written scope
What the automation does, what it must never do, what it may spend. Fixed price within two working days.
02
Built in the open
A staging instance in the first week, updated as I work, with a short written update every week.
03
The AI security review, before it goes live
Prompt injection, tool permissions, data leakage and exposed keys — the same review I sell standalone, run against what I built.
04
Handover and 30 days of fixes
Deployed on your accounts with written docs and a call. Then 30 days of fixes on anything in scope; new steps quoted in writing.
what you receive
A finding, as it arrives.
Every finding in a report says what the issue is, what an attacker gains, exactly how to reproduce it, and how to fix it — ranked by real-world impact, never a scanner result passed through. This one is from the public sample; the whole 12-page report is linked under it.
IDOR on order endpoint exposes any customer's orders
Order API · CWE-639 · OWASP API1:2023
What the issue is
GET /v2/orders/{orderId} verifies that the caller is authenticated but never checks that the order belongs to them. Order IDs are sequential integers, so the entire order table can be walked by incrementing a number.
What an attacker gains
Any authenticated customer can read every order in the system: line items, totals, delivery addresses, phone numbers and email addresses. For a retailer this is a personal-data breach affecting the whole customer base, reachable from a free signup.
Reproduction
# authenticated as customer A (own order)
GET /v2/orders/10442 → 200 OK // expected
# same session, another customer's order
GET /v2/orders/10441 → 200 OK
{"orderId":10441,"customer":{"name":"[REDACTED]", …
A short script walking IDs 10000–10500 returned 500 of 501 records. Enumeration was stopped at that point and no data was retained; the redactions above are applied at capture.
How to fix it
Enforce ownership server-side on every object access — the query itself should be scoped to the caller, not filtered afterwards. Replace sequential identifiers with UUIDs so that enumeration is not trivially possible even if a check is missed in future.
-- scope the query, do not check after fetching
SELECT * FROM orders WHERE id = $1 AND customer_id = $2;
Add an automated test asserting that customer A receives 404 for customer B's order. This class of bug returns whenever a new endpoint is added, so the guard needs to be a test, not a code review habit.
Sample report · fictional target · every finding, payload and identifier in it is synthetic. Real reports go to the client and nowhere else.
price and timeline — fixed in writing before it starts
1,2000
USD, from · ₹1,00,000 · one workflow or agent
20
working days from the call to a written scope and price
10
AI security review on every delivery, included
30 days0 days
of in-scope fixes after handover
“From” because a single workflow and a multi-agent system are different weeks. Reviews are invoiced at the start; builds 40% to start, 30% mid-way, 30% at launch. No hourly billing.
before you ask
Questions people actually ask.
Which tools do you use?+
n8n for workflows, the OpenAI and Anthropic APIs for the model, Supabase and Postgres for data, Vercel and AWS to run it. If you already use something else — Make, Zapier, a queue you own — I'll say on the call whether I'll work in it.
Whose API keys and accounts?+
Yours, from day one. The workflow runs in your n8n, your keys sit in your credential store, and you can revoke my access the day we hand over.
What does “secured” actually mean?+
The AI security review runs against the finished automation before it goes live: what the model can be talked into, what its tools can reach, what leaks, where the keys are. Findings in writing, fixed before launch.
Can you automate something that already exists?+
Yes — and if it currently lives on someone else's instance or key, moving it onto your own accounts with rotated credentials is usually the first job.
start
Tell me the job nobody wants to do by hand.
One sentence on what comes in and what should happen. I reply within one working day.
