agentleFS
NickKostovnyNickKostovny/marketing-playbooks/skills/defensibility-check/SKILL.md

defensibility-check

Skill0 starsChanged 2 months agoMITView on GitHub

Apply this skill whenever Nick is reviewing marketing content — his own drafts, LLM output, contractor work, agency copy, teammate submissions, or any written piece that will represent Invert externally. Trigger on signals like "review this," "does this work," "is this good," "what do you think of this copy," "tighten this up," or any moment where a piece of marketing language is being evaluated for quality. Also trigger when Nick shares a sentence or headline and asks whether it lands. The skill's job is to train Nick in a single reviewing move — the defensibility question — not to do the reviewing for him. Do NOT trigger for strategy, structure, or positioning questions; this is specifically about sentence-level language quality.

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 11 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈11

    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 the answer to that question requires a new abstraction — another word like "throughout," "across," "holistically," "compounding" — the …

    line 14
  2. Distinctive:

    When Nick shares content for review, do not simply critique it.

    line 28
  3. Distinctive:

    Read it back to him.

    line 32
  4. Distinctive:

    Ask: "What does this specifically mean?"

    line 33
  5. Distinctive:

    If the answer requires abstraction-swapping, flag the failure and propose a concrete rewrite.

    line 34
  6. Distinctive:

    If the answer points at something specific, confirm it passes.

    line 35
  7. Distinctive:

    Do this even when the sentence seems fine at first glance.

    line 37
  8. Distinctive:

    Do not rewrite silently. Nick needs to see the failing sentence and the rewrite side by side, so he learns the pattern.

    line 49
  9. Distinctive:

    Do not score content holistically ("this is 7/10").

    line 50
  10. Distinctive:

    Do not apply the test to every sentence.

    line 51
  11. Distinctive:

    Do not let Nick off the hook when his own writing fails the test.

    line 52

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

The Defensibility Check

The move

For every claim-making sentence in a piece of content, ask one question:

"What specifically do you mean?"

If the answer to that question requires a new abstraction — another word like "throughout," "across," "holistically," "compounding" — the sentence has failed. The language is doing the work that evidence should be doing.

A defensible sentence survives the question by pointing at something concrete: a person, an action, a specific thing, a measurable change.

Why this matters for Nick

Experienced reviewers spot undefendable marketing language instantly because it can't survive a "what do you mean?" question. "Scale up many different parts throughout the process" fails — it's a verb being used as a value prop with no defensible specifics underneath.

This is the skill that separates marketing that practitioners trust from marketing they dismiss. Nick's ICP (bioprocessing scientists, process engineers, R&D digital leaders) has been burned by vendor language for years. They read undefendable claims as either naïve or dishonest.

Building this muscle in Nick's reviewing practice means every piece of content — from any source — gets filtered through the same test before it goes out.

How to apply it

When Nick shares content for review, do not simply critique it. Walk through it with him, applying the move out loud so he can internalize the pattern.

For each claim-making sentence:

  1. Read it back to him.
  2. Ask: "What does this specifically mean?"
  3. If the answer requires abstraction-swapping, flag the failure and propose a concrete rewrite.
  4. If the answer points at something specific, confirm it passes.

Do this even when the sentence seems fine at first glance. The goal is repetition — Nick should start anticipating the question before you ask it.

Failure patterns to watch for

  • Verb-as-noun claims: "scale up," "optimize," "accelerate" used as standalone value props with no object.
  • Abstract compound nouns: "compounding insights," "seamless integration," "end-to-end visibility." These hide who, what, when.
  • Actor-less verbs: "innovation happens," "value is unlocked," "insights emerge." Rewrite to name the person doing the thing.
  • Range-words doing specificity work: "throughout," "across," "holistic," "comprehensive." These substitute scope for substance.
  • Upgrade vocabulary: language that sounds more sophisticated than the underlying idea warrants. The McKinsey-deck test — if it would appear in one without modification, rewrite it.

What to resist

  • Do not rewrite silently. Nick needs to see the failing sentence and the rewrite side by side, so he learns the pattern.
  • Do not score content holistically ("this is 7/10"). Go sentence by sentence on the claim-making ones.
  • Do not apply the test to every sentence. Transitions, connective tissue, and setup don't make claims. The test applies specifically to sentences that assert value or argue a point.
  • Do not let Nick off the hook when his own writing fails the test. The skill is worthless if it only catches other people's failures.

The self-diagnostic

The real goal is for Nick to stop needing this skill.

When Nick has internalized the move, he will start asking "what specifically do I mean?" while writing, which will change his drafts before they ever get reviewed. That's the endpoint — the skill makes itself progressively less necessary as the reviewing instinct becomes the writing instinct.

Track this. If Nick's drafts over time contain fewer flaggable sentences, the skill is working. If reviewer pushback on language decreases, even better. That's the real outcome measure.

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.