agentleFS
randommoniclerandommonicle/claude-skills/substantiate-outward-claims/SKILL.md

substantiate-outward-claims

Skill27 starsChanged 20 days agoApache-2.0View on GitHub

Any certification, award, statistic, or third-party performance claim shown to a customer needs a primary or independent source, or explicit attribution ("the manufacturer states…") — never asserted as fact, and never trusted from an AI summary, which blends marketing with real product names. Triggers when writing customer-facing copy, marketing pages, or an assistant's knowledge base entries that assert such claims. Does not fire on internal docs citing code or history — verified-citations owns those.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    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:

    Never launder a claim through an AI summary: Search-engine AI overviews and model memory blend marketing copy with real product data and …

    line 21
  2. Distinctive:

    A qualified claim is the contradiction restated: When two passages contradict, remove the claim and find the frame both can share.

    line 46

≈ “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

Substantiate outward claims

A public site nearly shipped "WoolSafe approved", a decibel figure, and a machine model designation, none of which could be substantiated — one of them contradicted by the manufacturer's own documentation while a search engine's AI overview confidently asserted the opposite. The exposure on outward claims is advertising standards and reputation, which is why this is separate from verified-citations (internal references).

The rules

  • Every outward claim gets a primary or independent source, or explicit attribution. "The manufacturer states a 76dB rating" is honest; bare "76dB" is your claim now.
  • An in-house assessment is not an independent certification. "Safe for wool" (our judgement) and "WoolSafe approved" (a certification body's mark) are different claims with different legal weight; never upgrade one into the other.
  • Never launder a claim through an AI summary. Search-engine AI overviews and model memory blend marketing copy with real product data and assert the blend confidently — in the evidence, confidently wrong. Chase the vendor's own documentation; in a vendor-dominated domain, the SEO layer repeats numbers with no methodology while the vendor publishes the actual behaviour.
  • A number with no traceable methodology is fabricated until shown otherwise — the archetype being a "75% of CVs auto-rejected" statistic that traced to a dead company's 2012 marketing.
  • Label evidence strength in research that feeds customer-facing content: well-evidenced, weakly evidenced, or myth — and say which.

What this skill does not do

It does not govern internal engineering citations (verified-citations) or the claims inside generated client documents (deliverable-integrity handles claim drift in rewrites). It governs what the outside world is told as fact.

Why

Outward claims outlive their authors: they get quoted, cached, and screenshotted, and an indefensible one is a legal exposure that no later correction fully retrieves. Evidence: ICC LESSONS_LEARNED L-009 with both addenda; this repo's LESSONS_LEARNED lesson 4.

Additions from the misses-log sweep (2026-10-22)

  • A qualified claim is the contradiction restated. When two passages contradict, remove the claim and find the frame both can share; appending a qualification keeps the contradiction. (RMC guide, 2026-08-11.)

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.