Showroom Cases / End-to-End Execution Examples

Typical problem

• You already have Python scripts, but first-time users do not know where to start.
• Many requests need an explanation of inputs, outputs, and steps before real execution begins.
• You need a low-friction experience entry instead of asking customers to connect a real environment immediately.

How ExecFabric handles it

• The user enters one task topic on the experience page, such as "organize the downloads directory."
• The platform calls task_brief_demo.py to generate a smallest pre-execution brief.
• The result returns the suggested execution chain, input preparation, and next-step entry point.
• This experience chain can later be replaced by the tenant's own real Python scripts.

Governance points

• At the experience stage it generates only a brief and does not write real business data directly.
• The current topic, execution time, and output content can still be recorded to form a smallest review chain.
• It is suitable for proving that "AI catches the request and then schedules a Python script" is already real.

Final deliverables

• A structured task-brief text.
• Suggested execution steps plus input and output guidance.
• A next-step entry for signup or for connecting real scripts later.

Suggested screenshot copy

• Main title: enter one task topic and generate a Python task brief.
• Subtitle: show how the no-login AI experience catches the request first and then schedules a Python script to output a pre-execution brief.
• Badge copy: lightweight experience / read-only preview / ready to connect formal scripts later.

Suggested demo script

1. I am not connecting a real business environment yet. This shows only the smallest execution chain first so you can understand how ExecFabric catches a request.
2. I enter one task topic, such as "organize the downloads directory." The platform first calls a Python script to generate a pre-execution brief instead of making a dangerous direct action.
3. The value of this step is that it makes the inputs, steps, and next entry clear before replacing it with your own formal script later.

Typical problem

• INFO, DEBUG, WARN, and ERROR are mixed together in logs, and manual filtering wastes time.
• Many Shell scripts live only in terminals, so it is hard for others to see what they actually do.
• You need a safe preview scenario instead of asking someone to run a real cleanup action immediately.

How ExecFabric handles it

• The user pastes a mixed log segment directly into the experience page.
• The platform calls clean_log_demo.sh for a preview run.
• The result returns the preserved WARN and ERROR lines plus a row-count summary.
• This chain makes it visible that Shell scripts have already been brought under one unified entry.

Governance points

• The current case is a preview and does not modify real server log files.
• The input log text, output result, and kept row count can all be recorded.
• It is suitable for demonstrating that Shell assets can also be brought into a controlled execution entry.

Final deliverables

• A filtered preview of the important log lines.
• A WARN and ERROR count plus output summary.
• A reviewable execution result and smallest audit record.

Suggested screenshot copy

• Main title: paste a log segment and preview Shell log extraction.
• Subtitle: bring terminal-side Shell preprocessing into one unified entry while keeping only WARN and ERROR highlights.
• Badge copy: result preview / does not change real logs / reviewable later.

Suggested demo script

1. Many Shell scripts usually live only in terminals and other people cannot see the outcome, so here I built a visible preview entry.
2. I paste one mixed log segment. The platform calls a Shell script for preview mode and extracts only WARN and ERROR instead of changing any real file.
3. The point is to show that Shell assets have already become controlled, demo-ready, unified capability units.

Typical problem

• You receive a CSV or Excel file but cannot even see the fields and missing values clearly at first glance.
• Every first pass still depends on manually opening the file and doing basic exploration again.
• You need the simplest file-driven sample to prove that the platform does more than run text-only scripts.

How ExecFabric handles it

• Upload one CSV or Excel file from the unified entry.
• The platform calls csv_profile_demo to generate an analysis report.
• The report outputs summary information, field statistics, and a preview of the first rows.
• The final result returns as an Excel artifact that is suitable for download and review.

Governance points

• Limit input file type and size so arbitrary files are not pushed into the execution chain.
• Record the batch number, source filename, and generated artifact path.
• The current case performs read-only analysis and result generation without overwriting the original file.

Final deliverables

• A summary worksheet.
• A columns profile worksheet.
• A preview worksheet plus a download entry.

Suggested screenshot copy

• Main title: upload CSV and generate a field overview and profile report.
• Subtitle: show how a no-login file flow moves from uploaded input to field profiling, preview rows, and a downloadable artifact.
• Badge copy: real and clickable / downloadable report / original file stays untouched.

Suggested demo script

1. The first two cases use text input. This case exists specifically to prove that ExecFabric can also handle file-driven flows.
2. I upload a CSV or Excel file. The platform first produces field profiling and a preview of the first rows, then generates a downloadable xlsx report.
3. For customers, this proves that the platform already has the smallest complete loop of upload input - execute processor - return result file.