triangulate

star 16

Verifies information through multiple independent sources with transparent confidence scoring. Classifies claim type (fact/opinion/prediction), shows where confidence comes from, flags bias. Use when: (1) "Verify this", (2) "Triangulate", (3) "Confirm from different sources". Triggers: "/triangulate", "verify this", "triangulate", "confirm from multiple sources".

vasilievyakov By vasilievyakov schedule Updated 2/21/2026

name: triangulate description: | Verifies information through multiple independent sources with transparent confidence scoring. Classifies claim type (fact/opinion/prediction), shows where confidence comes from, flags bias. Use when: (1) "Verify this", (2) "Triangulate", (3) "Confirm from different sources". Triggers: "/triangulate", "verify this", "triangulate", "confirm from multiple sources".

Triangulate

Verification of claims through 3+ independent sources. Transparency: show sources, weights, where confidence comes from.

Workflow

  1. Classify the claim type
  2. Search in parallel using available web search tools
  3. Score sources (type, weight, independence, date, bias)
  4. Calculate confidence
  5. Format response according to claim type

Claim Classification

Type Indicators Verification goal
FACT Dates, numbers, events, tech specs True or false
OPINION "Better", "worse", evaluations, comparisons Map of positions
PREDICTION "Will be", "will become", about the future How well-founded
MIXED Contains multiple types Decompose

Strategies by Type

FACT

Search: official sources, documentation, peer-reviewed, primary data.

Verdicts:

  • CONFIRMED: fact is true
  • REFUTED: fact is false
  • PARTIAL: partly true
  • INCONCLUSIVE: insufficient data

OPINION

Search: different viewpoints, arguments from each side, who agrees/disagrees.

Verdicts:

  • CONSENSUS: most agree
  • MAJORITY X: most support X, opponents exist
  • DIVIDED: opinions roughly equal
  • MARGINAL: this is a minority opinion

Output: opinion map (FOR / AGAINST / NEUTRAL) with percentages and arguments.


PREDICTION

Search: author's track record, methodology, expert consensus, alternative scenarios.

Verdicts:

  • WELL-FOUNDED: methodology and track record exist
  • PARTIALLY FOUNDED: arguments and counterarguments both present
  • SPECULATIVE: weak methodology
  • CONTRADICTS CONSENSUS: experts think otherwise

Output: assessment of author, methodology, assumptions.


MIXED

  1. Split into parts by type
  2. Verify each part with its own strategy
  3. Synthesize overall conclusion

Source Hierarchy

Weight Source types
High Official docs, peer-reviewed, government sources, primary data
Medium Tech publications, experts, educational resources, Wikipedia with sources
Low Blogs, forums, social media, anonymous

Confidence Score

Confidence = Quality (40%) + Independence (30%) + Consensus (20%) + Freshness (10%)

Show the user what the score is composed of.


Warnings

Flag when detected:

Signal When
BIAS All sources lean one way
ECHO Sources cite each other (false triangulation)
OUTDATED Information may have changed
INSUFFICIENT Too few sources for triangulation

Principles

  • Search for refutations, not only confirmations
  • Verify source independence
  • Official documentation > blogs
  • Acknowledge uncertainty honestly
  • Note publication dates
Install via CLI
npx skills add https://github.com/vasilievyakov/miracle-infrastructure --skill triangulate
Repository Details
star Stars 16
call_split Forks 1
navigation Branch main
article Path SKILL.md
More from Creator
vasilievyakov
vasilievyakov Explore all skills →