k-deep-research

Deep research requires multi-source triangulation to gather, score, and synthesize evidence. This Skill collects and compares sources, applies credibility scoring on a 0-10 scale, and preserves uncertainty by highlighting contradictions and gaps rather than forcing a quick answer.

Source credibility scoring rates every source on merit using an explicit 0-10 scale, not prestige. For contested topics, it evaluates evidence quality across obfuscated sources, enabling pattern recognition and adversarial analysis to detect narrative shifts and suppression indicators.

To generate an Obsidian-ready Markdown report, the Skill synthesizes gathered evidence with YAML frontmatter. It outputs cited reports including source credibility scores, identified contradictions, open questions, and confidence levels after completing triangulation and adversarial analysis.

Yes, curl and wget are required dependencies for gathering evidence. They enable source loading from obfuscated sources, supporting technical analysis and institutional research by fetching data needed for credibility evaluation, pattern recognition, and triangulation.

Adversarial analysis and pattern recognition are warranted when investigating contested topics requiring triangulation. They detect timing patterns, coordination signals, suppression indicators, and narrative shifts, ensuring rigorous investigations handle explicit uncertainty rather than stopping at surface-level summaries.

The best way to investigate open-ended questions is systematic evidence gathering with multi-source triangulation. This approach applies credibility evaluation, pattern recognition, and adversarial analysis when relevant, producing cited reports with contradictions, gaps, and confidence levels clearly stated.