SuperAPI AI Directory

credible-claims-and-safe-authority

General & other

View repository

Updated 10-15 23:40

About this skill

Credible Claims and Safe Authority Skill

Purpose

Use authority, numbers, and institutional credibility without exaggerating or creating legal, ethical, or reputational risk.

Claim Safety Ladder

From safest to riskiest:

  1. Direct factual claim
    • "Built and deployed X."
  2. Quantified but bounded claim
    • "Surfaced 770 addressable hours."
  3. Shared-outcome claim
    • "Helped the team maintain velocity."
  4. Inference claim
    • "This suggests a higher-leverage roadmap."
  5. Borrowed authority claim
    • "Adapted from public guidance by X."
  6. Endorsement-like claim
    • "X recommends this." Only use if explicitly true.

Use Careful Language For

Addressable Opportunity vs Guaranteed Savings

Say:

surfaced 770 addressable hours

Avoid:

saved 770 hours

unless the hours were actually saved.

Team Outcomes

Say:

helped maintain team velocity

Avoid:

single-handedly maintained team velocity

unless that is provably true.

Institutional Authority

Say:

at a World Economic Forum-recognized AI scale-up

Avoid:

World Economic Forum-backed

unless backed is accurate.

Inspired by Credible People

Say:

adapted from practices learned from a senior engineering lead

Avoid:

Google's official method

unless it is official.

Quote Usage

When using quotes:

  • Keep excerpts short.
  • Use brackets [...] to omit irrelevant parts.
  • Preserve meaning.
  • Attribute clearly.
  • Do not make a quote sound broader than it is.

Checklist

  • Is every number explainable?
  • Could a former colleague agree the claim is fair?
  • Is borrowed authority framed as origin/context, not endorsement?
  • Are "helped," "solo-built," and "led" used precisely?
  • Does the claim stay impressive without becoming misleading?