analyze-argumentation-flow — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited analyze-argumentation-flow (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Read whole text once before flagging anything. Reconstruct central thesis in one sentence. Do this even when author never states it. Report missing thesis as finding when you cannot reconstruct it.
List supporting claims in order and note evidence for each. Judge argument author makes, not one you wish they made. Distinguish flaws of reasoning from flaws of taste. Distinguish the conclusion is wrong from the argument is weak. Report only second kind.
Flag every factual claim without support. Surface unstated premises argument leans on. Check every key term for equivocation.
Test inference at every therefore, so, thus, or hence. Test it at every which means or as a result too. Hunt for circular reasoning. Check each premise's independent believability.
Name specific fallacies and quote triggering line. Use names from references/fallacies.md. Treat one anecdote as one anecdote. Watch for cherry-picked evidence and survivorship bias. Mark every analogy and ask where it breaks.
Flag rhetorical questions that stand in for arguments. Treat obviously, clearly, and everyone knows as red flags. Treat it goes without saying and as we all agree same way. Watch for loaded language replacing evidence.
Check author addresses strongest counterargument. Use one fair opponent would raise. Notice when burden of proof shifts onto reader.
Hunt for contradictions across paragraphs. Compare introduction's promise against conclusion's delivery. Flag paragraphs that restate earlier claims. Mark restatements that add no evidence, distinction, or consequence. Notice scope creep, where narrow claim inflates into sweeping one. Check order of reasoning. Distinguish description from prescription.
Forgive slips that leave conclusion intact and challenge load-bearing ones. Raise only objections honest reader would make.
Report findings as numbered list. Give quoted snippet and fallacy name in each item. Add one-sentence explanation and one-line remedy. End with one-paragraph verdict. Name which load-bearing claims survive and whether thesis stands.
Leave prose and author's voice intact. Keep author's conclusion as theirs.
Confirm each finding carries snippet, fallacy name, explanation, and remedy. Confirm report ends with one-paragraph verdict naming which claims survive.
Take this argument as input.
Sales rose after the rebrand, so the rebrand drove the growth.Return this report.
1. Post hoc fallacy — "drove the growth" treats sequence as cause.
Explanation: rebrand and growth could share another driver.
Remedy: compare against comparable period without rebrand.
Verdict: causal claim unsupported, so thesis does not stand.~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.