agentleFS
forefyforefy/.context/skills/generic-utils/variant-table/SKILL.md

variant-table

Skill153 starsChanged 44 days agoMITView on GitHub

Lay draft copy out as a table with three distinct rewrites per line and a top pick. Use to compare wording for posts, headlines, CTAs, or microcopy.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 5 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If two variations could be swapped without a reader noticing a shift in strategy, they are too similar - rewrite one.

    line 34
  2. Distinctive:

    When unsure of granularity, match whatever unit the user referenced last, or ask if it's ambiguous.

    line 44
  3. Distinctive:

    Keep every cell tight. A variant is a candidate line, not a paragraph of explanation.

    line 49
  4. Distinctive:

    Do not pad to three if the unit is trivial (e.g.

    line 50
  5. Distinctive:

    After the table, offer to stitch the picks (or a chosen column) into a clean final draft.

    line 51

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Variant Table

Turn a piece of draft copy into a side-by-side comparison table so the user can compare wording options per unit and pick.

When to use

Any time the user is refining wording and would benefit from seeing distinct options laid out rather than a single rewrite. Works for any content type: social posts, headlines, taglines, email lines, bios, CTAs, microcopy, doc sentences, spec bullets, commit messages. Content-agnostic - never assume it is a LinkedIn post.

Output format

A markdown table. One row per logical unit of the source. Columns:

| # | Current | A - <angle> | B - <angle> | C - <angle> | My pick |

  • # - unit index plus a 1-2 word role label in parens, e.g. 1 (hook), 4 (CTA), 2 (bullet). The label names the unit's job, inferred from the content.
  • Current - the user's existing text for that unit, verbatim. If the user gave no existing draft, leave a short placeholder or omit this column.
  • A / B / C - three rewrites. See the distinctness rule below. Rename the angle after the dash per row to whatever actually differs (e.g. A - blunt, B - contrarian, C - technical).
  • My pick - one recommendation for that row, in bold. It may be one of A/B/C, the current text ("keep"), or a fresh hybrid. Add a terse parenthetical reason only when it aids the choice.

The distinctness rule (most important)

The three variations must differ in approach, not just word swaps. Before writing a row, pick three different axes and commit one variation to each. Draw from axes like:

  • tone: blunt / playful / formal / contrarian
  • structure: statement / question / fragment / story-fragment
  • angle: pain-first / benefit-first / curiosity-gap / social-proof / technical-precision
  • length: terse vs flowing

If two variations could be swapped without a reader noticing a shift in strategy, they are too similar - rewrite one. Label each column with the axis it took so the contrast is legible.

How to segment

Split the source into the units the user is actually deciding on:

  • prose → paragraphs (or sentences if they're iterating sentence-level)
  • list → bullets
  • headline + subhead → separate rows
  • table/form → cells or fields

When unsure of granularity, match whatever unit the user referenced last, or ask if it's ambiguous.

Style

  • Match the user's established preferences (terse, no em dashes in the copy itself, etc.). The - in column headers is a label separator, fine to keep, but keep it out of the generated copy if the user dislikes it.
  • Keep every cell tight. A variant is a candidate line, not a paragraph of explanation.
  • Do not pad to three if the unit is trivial (e.g. a URL); mark it "keep".
  • After the table, offer to stitch the picks (or a chosen column) into a clean final draft.

Example shape

#CurrentA - bluntB - contrarianC - curiosityMy pick
1 (hook)We help teams ship faster.Ship faster. Full stop.Most "ship faster" tools slow you down.What if your slowest step wasn't code?Most "ship faster" tools slow you down.

Then: "Say the word and I'll stitch the pick column into a final draft."

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.