# Agent Guardrail Desk

> Agent Guardrail Desk is an AI agent guardrails review for agent builders, AI-built app makers and platform teams in the United States: it tests six guardrails on one AI agent or automation before it runs unattended, for a fixed $1,500. Mike Rodgers runs every review remotely from Denver, Colorado.

## Offer and prices
- Pre-Flight Review: $1,500 once, per agent or automation. Six guardrails tested on a staging copy, the Pre-Flight Report, a ranked fix list, two configs and a 30-minute readout. Delivered in 5 business days.
- Guardrail Fix Sprint: Quoted in writing, fixed price, quoted after the review. We install every guardrail that failed and rerun each acceptance test with your team watching. The fixed price is quoted in writing before any work starts, and the $1,500 review fee is credited in full if you book within 30 days.
- Guardrail Recheck: $400 a month, per agent, month to month. The six acceptance tests rerun after each release (up to two a month) and at least once a month, with a one-page log. Cancel with 30 days notice.

## The six guardrails (each with a pass or fail acceptance test on a staging copy)
- Spend Cap (prevents: the runaway bill): We plant a retry loop on staging. Pass: the run stops at the cap and the alert reaches the named person.
- Kill Switch (prevents: the agent nobody can stop): We pull the switch during a live staging run. Pass: the run in flight stops, no new run starts, and the time to stop is written down.
- Permission Scope (prevents: the write it should never make): We ask the agent to write outside its scope and search the frontend bundle for keys. Pass: the write is refused and no secret ships to the browser.
- Silent-Break Watch (prevents: the run that keeps going on empty): We rename one form field on staging. Pass: the run halts and alerts before it writes an empty record.
- Hijack Tests (prevents: the instruction hidden in a ticket or a web page): We plant an instruction inside a document the agent reads. Pass: the agent refuses it, or stops for human approval before any destructive step.
- Run Log (prevents: the run nobody can explain): We pick one run from the past day. Pass: those four questions are answered from the log in under five minutes.

## Who it is for
- Pre-Flight for Agent Builders: Engineers shipping tool-using agents, multi-agent pipelines and MCP servers.
- Pre-Flight for AI-Built Apps: Founders and makers shipping apps built with Lovable, Bolt, Cursor or Claude Code on Supabase.
- Pre-Flight for No-Code Automations: Ops owners running Zapier, Make or n8n flows from forms and webhooks into a CRM.
- Pre-Flight for Platform Teams: Platform, security and ops leads whose internal agents can write to CRM, ticketing and docs.

## Boundaries (what this is not)
Not included: production access, a compliance certification (SOC 2, ISO 27001), a full penetration test, a hosted monitoring product, model training, or any promise that an agent will never fail. The boundary is part of the product.

## Accountability
Named owner of mistakes: Mike Rodgers runs every review and signs every report.
When a guardrail fails in the review, it is marked FAIL with the evidence and the fix, never softened. After a Fix Sprint, if a guardrail we installed fails its own acceptance test on the same setup within 30 days, Mike fixes it again at no charge, and the failure is written into the run log you keep.

## Pages
- [Home](https://agent-guardrail-desk.vercel.app/): the offer, the six guardrails, prices, rulebook, FAQ and request form
- [Pre-Flight for Agent Builders](https://agent-guardrail-desk.vercel.app/agent-builders/): Engineers shipping tool-using agents, multi-agent pipelines and MCP servers
- [Pre-Flight for AI-Built Apps](https://agent-guardrail-desk.vercel.app/ai-built-apps/): Founders and makers shipping apps built with Lovable, Bolt, Cursor or Claude Code on Supabase
- [Pre-Flight for No-Code Automations](https://agent-guardrail-desk.vercel.app/no-code-automations/): Ops owners running Zapier, Make or n8n flows from forms and webhooks into a CRM
- [Pre-Flight for Platform Teams](https://agent-guardrail-desk.vercel.app/platform-teams/): Platform, security and ops leads whose internal agents can write to CRM, ticketing and docs
- [How do I cap what an AI agent can spend?](https://agent-guardrail-desk.vercel.app/cap-ai-agent-spend/): Where a spend cap has to live, what it counts, and the test that proves it stops a loop.
- [What should an AI agent kill switch actually do?](https://agent-guardrail-desk.vercel.app/ai-agent-kill-switch/): The four things a kill switch must stop, who holds it, and how to prove it works.
- [How do I catch a renamed form field before it writes empty records?](https://agent-guardrail-desk.vercel.app/silent-automation-failures/): Silent breaks in automations: what to check before each write, and the test that proves it.
- [How do I check an AI-built app is secure before real users?](https://agent-guardrail-desk.vercel.app/ai-built-app-security-check/): The pre-launch routine for Lovable, Bolt, Cursor and Claude Code apps, and what to test after it.
- [The Pre-Flight Review: price, scope and timeline](https://agent-guardrail-desk.vercel.app/pre-flight-review/): $1,500 per agent, report in 5 business days, credited in full against a Fix Sprint within 30 days.
- [See a sample Pre-Flight Report](https://agent-guardrail-desk.vercel.app/sample-report/): The report format, filled in for a placeholder agent. Demonstration, not a client result.

## Contact
Request form: https://agent-guardrail-desk.vercel.app/#form
Email: mike@rodgersintelligence.com
Location: Denver, Colorado; remote service across the United States

## Operator
Agent Guardrail Desk is a Rodgers Intelligence Group LLC company. 1767 North Pearl Street #304, Denver, CO 80203
Mike Rodgers runs every review.

Last updated: 2026-09-26