Modular Code Architect
Forces a plan before any code lands. Code ships one file at a time, each under a strict size budget, so features stay reviewable and refactors stop drifting into rewrites.
DownloadResources
Skills and prompts from production workflows. Ready to copy and use. These resources are a preview of what ElleMentys will offer: practical AI education for beginners, built around real tools and real workflows. No coding required, no jargon, just clear steps anyone can follow.
Forces a plan before any code lands. Code ships one file at a time, each under a strict size budget, so features stay reviewable and refactors stop drifting into rewrites.
DownloadRuns a 6-layer review after every file: security, correctness, coherence, usability, performance, style. A red flag halts the build; warnings get logged and fixed before the next file.
DownloadScopes a topic, runs structured web verification, and filters contradictions from confirmations. Outputs a short report with dated sources so a decision can be made without a second tab.
DownloadAuto-briefs at the start of any conversation with 2-3 targeted web searches around the task. Ten seconds of setup saves twenty minutes of chasing stale context.
DownloadTurns every correction into a durable rule. The same mistake never ships twice because it moves straight into the ruleset the next session reads first.
DownloadAct as a Make.com architect. I will describe a workflow in plain English. Return a module-by-module plan with trigger, mapper expressions, error branches, and cost per run. Flag anything that needs a webhook secret or rate limit before I build it. If the workflow connects to Notion, Slack, or email, specify the exact module name and required fields. Ask one clarifying question if the request is ambiguous.
When:Use when scoping a new Make.com scenario and you want an executable plan, not a pitch.
I will paste an API error response and the request that triggered it. Identify whether the issue is authentication, rate limiting, schema mismatch, timeout, or server-side. Explain the root cause in one sentence. Give me the exact fix for the request. If the fix requires a config change, show the before and after. If this is a webhook or Make.com HTTP module, include the corrected JSON body.
When:Use when an API call fails in production and the error message is not clear enough to act on immediately.
Compare these two prompt versions. Run them mentally against 5 realistic inputs I will provide. For each input, predict the output structure, flag regressions, and rate the change as safer, neutral, or risky to ship. If the prompt feeds into an automation (Make.com, Notion, or Telegram), check whether the output format still matches the downstream parser. End with a single recommendation.
When:Use before replacing a production prompt so you do not ship a silent regression.
Given this 90-day usage trend and support ticket list, score churn risk on a 1-5 scale. Name the 3 signals that drove the score, one leading indicator to watch next week, and one retention play with a specific owner. If you have access to a CRM or Notion database, suggest which fields to update with the risk score. Do not pad the answer.
When:Use during weekly book-of-business review to surface accounts that need intervention.
Rewrite this renewal email so it sounds like a human who actually knows the account. Keep it under 120 words. Lead with a specific outcome from the last quarter, name the renewal date, and end with one concrete next step. If account notes are in Notion or a CRM, reference a real data point instead of a generic placeholder. Cut every filler phrase.
When:Use when a templated renewal draft feels generic and needs sharpening before send.
I will give you a topic. Produce 5 opening hooks for a LinkedIn post. Each under 15 words. No buzzwords, no rhetorical questions. Voice: direct, short sentences, first person. If I provide a Notion page or draft, pull the strongest angle from the content instead of guessing. Flag the hook you would lead with and say why in one line.
When:Use at the start of a writing session to skip the blank page and pick an angle fast.
I will describe a manual task I want to automate. Estimate how many minutes per week I spend on it, calculate the yearly time cost, and compare it against the estimated build time and monthly tool cost for Make.com or a similar platform. Tell me if the automation pays for itself in under 3 months, under 6 months, or not worth building. If not worth it, suggest a simpler shortcut using existing tools like Notion templates or calendar reminders.
When:Use before committing time to build an automation so you do not spend 8 hours saving 5 minutes a week.
I will paste raw meeting notes or a transcript. Extract every action item with its owner and deadline. Flag any decision that was made. List any open question that was not resolved. Format the output so I can paste it directly into a follow-up message or a Notion page. Keep it under 10 lines.
When:Use right after a meeting to send a clean follow-up before the context fades.