fact-checking
/fact-checking
Verify claims before stating them: trace to the source, corroborate independently, and separate fact from inference. Use when accuracy matters, especially before an agent asserts facts, numbers, or quotes.
- Author
- Amey-Thakur ↗
- Repository
- AI-SKILLS ↗
- Licence
- MIT
- Last repository change
- 05.09.2026
Assessment and security
- Files in this package
- 1
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 05.09.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Verify claims before stating them: trace to the source, corroborate independently, and separate fact from inference. Use when accuracy matters, especially before an agent asserts facts, numbers, or quotes.
Original author description
Verify claims before stating them: trace to the source, corroborate independently, and separate fact from inference. Use when accuracy matters, especially before an agent asserts facts, numbers, or quotes.
How it works
- Method
- Boundaries
SKILL.md preview
--- name: fact-checking description: "Verify claims before stating them: trace to the source, corroborate independently, and separate fact from inference. Use when accuracy matters, especially before an agent asserts facts, numbers, or quotes." --- # Fact-checking A confident statement is not a true one. Fact-checking is the discipline of verifying a claim before you assert it: finding where it actually comes from, confirming it independently, and being honest about what you could not confirm. It is the guard against passing on plausible falsehoods, including your own. ## Method 1. **Isolate the checkable claim.** Separate the specific factual assertion (a number, a date, a quote, an event, a causal claim) from the surrounding interpretation. Vague claims cannot be checked; pin down exactly what is being asserted before verifying it. 2. **Trace to the primary source.** Follow the claim back to its origin: the study, the official record, the actual quote in context, the raw data (see web-research, source-evaluation). A statistic repeated across the web often traces to one source that said something narrower, older, or different. The original is the check. 3. **Corroborate independently.** Confirm important claims across sources that do not depend on each other. Many pages agreeing means little if they all copy one origin; independent confirmation from separate evidence is what verification requires. One source is a lead, not a confirmed fact. 4. **Check quotes and numbers in context.** Quotes get truncated to reverse their meaning; statistics get stripped of their caveats, denominators, and dates. Verify the exact wording and the full context, not the paraphrase. A number without its units, base, and as-of date is not yet a fact. 5. **Distinguish fact, inference, and opinion.** Be clear about what is established (verified fact), what is reasonable interpretation (inference you are drawing), and what is contested or unknown. Present each as what it is; dressing an inference as a fact is a subtle form of getting it wrong. 6. **Flag what you could not verify.** When a claim cannot be confirmed, say so plainly rather than asserting it or quietly dropping it. "Reported by X but not independently confirmed" and "I could not verify this" are honest and useful; confident assertion of the unverified is the failure to avoid. ## Boundaries - Fact-checking scales with stakes and reach: a private low-stakes note needs a light touch; a public claim, a decision input, or anything an agent asserts to a user needs rigor, because a confident wrong answer does real harm. - For an AI agent, this is the guard against confident confabulation: verify facts, numbers, and quotes against real sources before stating them, and never invent a citation (see llm-guardrails, structured-output's grounding). - Verification reduces error; it cannot reach certainty on genuinely contested or unknowable questions. There, the honest output is the disagreement and the uncertainty, not a manufactured answer.
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · Amey-Thakur/AI-SKILLS | 90 | MIT |
| github/gh-aw | 100 | MIT |
| nimadorostkar/Claude-Skills-collection | 100 | MIT |
| axiomantic/spellbook | 100 | MIT |
| seb1n/awesome-ai-agent-skills | 100 | MIT |
| sokolmarek/researcher | 100 | MIT |
| rohitgehe05/mindpowers | 90 | MIT |
Database snapshot: 13.11.2026 · package version c6caa77ced2f (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100checking-facts
Checks factual assertions (numbers, quotes, dates, versions, API or tool behavior, what a source actually says) against primary sources read this pass, confirming each load-bearing fact by a second independent means. Use before a packet, doc, review, or release note relies on a… - 90rseng-fact-checking
Covers verifying facts and sources at the content level: checking that a cited source actually contains and supports the claim it is cited for (claim-source alignment), assessing source trustworthiness (peer-review status, venue reputation, predatory-publishing signals, primary… - 90verify-claims
Fact-checks a draft, answer, report or AI-generated text before it is used or published. Use when the user asks to check, verify or fact-check something; before finalizing anything with facts, numbers, quotes or citations; or for high-stakes content (legal, medical, financial,… - 100fact-check
Use when the user asks to fact-check a manuscript, verify or audit its citations, or check that cited papers support the claims attributed to them. Trigger phrases "fact-check", "verify citations", "audit citations", "check the sources", "do these papers actually say that",… - 100dissect
Investigate complex claims: decompose into atomic verifiable components, resolve vague entities, verify each independently, separate confirmed facts from narrative interpretation. Use to fact-check complex claims, investigate narratives mixing fact with interpretation, or when a… - 100fact-checker
Systematic fact verification and misinformation identification using evidence-based analysis. Use when: verifying claims, checking facts, identifying misinformation, evaluating source credibility, or when user asks to "fact check", "verify", "is this true", or mentions claims that need validation.