Hire an AI automation reliability reviewer

Everything it can hand back

Your task sets the actual outputs. The specialist picks the methods that cover them.

• A check of nine parts of the design, each marked sound or weak, with a fix for each weak one
• A repaired description of the automation
• A written operating procedure for people who run it
• A test checklist to try on a small real case first
• Recovery steps for a failed or doubled run

How it does the job well

• Repair only what is weak. Keep the intended outcome, scope and voice.
• Do not invent a tool, schedule, permission, number of items, test result or file. Use neutral wording ("the existing test") or write MISSING:.
• Do not search for other automations. Search only if the user names a published loop or asks for alternatives.
• Do not run, connect to, change or start the automation. Nothing here sends, edits or schedules anything.
• Keep what a plan says apart from what a test showed. No test result exists until the user runs the checklist and reports it.
• Clarify the purpose, audience and limits, such as length, format and deadline, first.
• Use the user's materials and voice. Ask for samples when tone matters.

Mistakes it avoids

• Generic output that ignores the user's details.
• Inventing facts, quotes or sources.
• Long answers when a short one was asked for.

What a good result looks like

An example, not your task: You built a Zapier, Make or n8n automation, or an AI loop, and want to check it before it touches real customers or money.

• Design check. Nine parts of the automation, each marked sound or weak, with a fix for each weak one.
• Repaired design. The automation as it reads after you accept the fixes, with the intended outcome unchanged.
• Operating procedure. Purpose, who does what, the steps and the exceptions, written for the person who runs it.
• Test checklist. Cases to try on one small real case before you rely on it, with the expected result for each.
• Recovery steps. What to do when a run fails, runs twice or stops halfway, and who does it.

How the work is checked

• All nine parts are in the design check, each with a reason.
• Duplicate runs, a half-finished run and who can stop it are each covered in the procedure or recovery steps.
• The checklist has an expected result for each case and leaves pass or fail blank.
• The repaired design keeps the user's intended outcome.
• States which checks passed, failed or could not run, and names missing inputs.

These checks have not been run against a model yet. Your AI uses them to check its own work.

Capabilities

• Repeatable AI loop designer. Audits the automation you pasted. Checks the trigger, scope, inputs, exceptions, saved state, duplicate protection, retries, approval points and stopping rules. Repairs only what is weak. Hands step 2 the repaired design. Check before moving on: every weak part has a fix or a reason it was left.
• Process and SOP writer. Writes the procedure from the repaired design. It has a purpose, who does what, the steps, the exceptions, a test checklist and recovery steps. Marks anything you did not tell it as MISSING. You run the checklist yourself before you rely on the procedure.

How the role works

1. Checks your task, context and available tools.
2. Picks the methods needed and prepares a work plan.
3. Carries out supported tasks, with each method's limits and review gates.
4. Checks the real deliverables and corrects supported faults.
5. Hands back files, evidence and unresolved gaps.

Your specialist's working setup

Download the base specialist without a task form or account. README.md and SETUP.md explain how to use the folder. SOUL.md defines its scope. INTAKE.md asks missing questions inside your AI. PLAYBOOK.md, SKILLS.md, BEST_PRACTICES.md, CHECKS.md and HANDOFF.md guide the work and its checked deliverables.

Complete associated skills, supporting files and licences are included. Choose only the methods needed for the actual task. Keep context, working files and accepted feedback in your AI project.

Use this specialist through MCP or the API

Call prepare_specialist with role_id check-a-new-automation-before-relying-on-it only. No task is required to load it. Your AI asks what job you need done, what you want back and which relevant context it can use.

For task-led hiring, call find_specialist, explain the suggested specialists and let the user accept or change the team.

Complete ZIP: Download base specialist. API: GET /api/v1/specialists/check-a-new-automation-before-relying-on-it/pack. The host AI supplies its own model, tools and execution; a pack does not connect accounts or run a task.

Inputs and helpful context

The specialist starts from the task brief. Only active methods require their inputs.

For the primary method, when used

• A description of the automation, or its prompt or settings pasted in
• What starts it, such as a form, a new row or a schedule
• The data fields it reads, and where the results go
• One small real case to test with, such as a sample order (optional)
• Who approves, and who to call when it fails (optional)

Helpful context (optional)

• A log or screenshot of a run that went wrong, so the check can match a real failure.
• A named person who approves changes.

Tools and access

• Claude, ChatGPT or Gemini: Any of them can audit a pasted description. Use one that opens files if your automation is saved as a file.
• The tool the automation runs in (optional): You run the test checklist there. This workflow does not connect to it or run anything.

Functions and deliverables

• Automation reliability reviewer: Review an automation's triggers, inputs and failure paths. Get a process map, test checklist and recovery guide.
  • A check of nine parts of the design, each marked sound or weak, with a fix for each weak one
  • A repaired description of the automation
  • A written operating procedure for people who run it
  • A test checklist to try on a small real case first
  • Recovery steps for a failed or doubled run
• Recurring AI task prompt builder: Turn a recurring task or existing loop into a bounded prompt with checks and stop rules.
  • Loop design with trigger and actions
  • Verification and stop rules
  • Guardrails and handoff points
  • Review notes on an existing loop
  Source: Alireza Rezvani.
• Process and SOP writer: Turn how your team works into a clear SOP. Give it a process name or old notes. You get a RACI table, a flow, steps and exceptions.
  • A step-by-step written procedure
  • A RACI table
  • A process flow
  • An exceptions and edge cases table
  • Metrics to track the process
  Source: Anthropic.
• Zapier and Make automation planner: Describe the trigger, apps and result. Compare Zapier and Make patterns, map the steps and plan tests, costs and failure handling.
  • A tool recommendation with reasons
  • A trigger, filter and action layout
  • Test and naming tips
  • A sign for when to switch to code
  Source: sickn33.
• me.md builder: Answer a few questions, or let your AI use what it remembers. Get one private file that tells any AI who you are.
  • One me.md file with up to ten dated sections
  • A list of facts taken from memory, for you to check
  • A short note of what it removed for privacy
  • The two next steps, to save it and set it up

Tasks you can use it for

• You built a Zapier, Make or n8n automation, or an AI loop, and want to check it before it touches real customers or money.
• An automation ran twice, skipped a record or stopped halfway, and you want to know why.
• You are handing an automation to a teammate and they need a procedure, not just your memory.
• You want a test checklist and recovery steps written down before the first live run.

Limits to check before starting

• You have no automation yet and want one built. Use the n8n workflow reliability planner workflow.
• You want to find which chores are worth automating. Use the Routine task automation planner.
• You want a plan for an AI agent with tools and permissions. Use the AI agent launch security reviewer workflow.
• You want the automation tested for you. This workflow checks the design and writes the test list. You still run the test.

Best way to use this role

Give it one clear outcome, the actual files and a definition of done. Keep the work in one AI project so it can reuse checked context and earlier files.

• Download the pack and give it to your AI, such as Claude, ChatGPT or Gemini.
• Paste the automation as it stands now: the prompt, the steps or the settings. Add what starts it, what it reads and where results go.
• Read the weak parts in step 1. Decide which fixes you accept. Say so before the AI writes the procedure.
• Run the test checklist yourself on one small real case. Mark each line pass or fail.
• Give the procedure to the people who will run it. Keep the failed checks with it.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/check-a-new-automation-before-relying-on-it for its role instructions and pinned sources.

This is a prepared AI role. No task has run. Your AI carries out authorized work using its available tools and reports the actual deliverables and checks.