Record the data-entry task once; SkillForge produces a SKILL.md an AI agent can repeat for every row.
Record a data-entry skill free What is SKILL.md?Already have a Chrome DevTools Recorder export or a Selenium IDE project? Convert it to an agent skill in your browser — free, no signup, nothing uploaded. You can also check an existing SKILL.md the same way.
Copy-pasting between a spreadsheet and a web form is a perfect agent task. Record it once and SkillForge extracts the steps into a SKILL.md your agent replays for each record.
Say you receive new leads in a Google Sheet and must enter each into your CRM. You'd normally do this by hand, row by row. Instead:
Because the steps are captured from a real demonstration, you don't describe selectors or write code — you just show the task once.
The skill you get describes one pass of the task. There is no loop in it, no row counter and no cursor — iterating is the agent's job, and it is worth being clear about what that means in practice before you point it at a thousand-row sheet.
See also support-macro automation, QA test automation, and SkillForge as an RPA alternative.
Every one of these produces the same artefact: a portable SKILL.md your assistant replays. The format is an open MIT spec, and if you already have a Chrome DevTools Recorder export or a Selenium IDE project you can convert it to a SKILL.md for free without recording anything again. Full index in the guides.
Want to see what you actually get? Download a complete SKILL.md — free, no account, generated from a 34-second screen recording. Hand it to your assistant, or paste it into the validator.
Can it handle many rows?
The SKILL.md describes the per-record steps; your agent loops it over each row of your data.
How does it know which spreadsheet column goes in which field?
When you record one row, SkillForge captures which value you typed into which field. You map those to your data columns, and the agent applies the same mapping to every row.
Do I need to write any code?
No. You record the task once; SkillForge writes the step spec. Your agent executes it.
What happens if it fails halfway through a thousand rows?
The agent decides, because the skill has no exception branch to tell it otherwise — it may skip the row, retry it, or stop. That is worth settling before an unattended run rather than during one: say explicitly in your instruction whether a row that cannot be entered should be skipped and logged or should halt the batch. And add a column the agent fills in as it goes, so your sheet records progress; that is the cheapest substitute for the run log we do not produce.
Does the skill loop over my rows by itself?
No. It describes one pass of the task — one row, start to finish. There is no loop, row counter or cursor in the file; iterating is something you ask the agent to do with the skill and your data. The practical consequence is that nothing in the file remembers where a stopped run got to.
What if the form changes?
Because an AI agent reads the screen and follows the SKILL.md steps, minor layout changes are usually handled gracefully — unlike selector-bound scripts. For big changes, just re-record.