Frame research prompts requiring triangulated evidence for ADR decisions.
npx skills add https://github.com/humansys/raise --skill rai-research-humansys
Please help me install this Agent Skill. Skill: rai-research Source: https://github.com/humansys/raise/tree/main/packages/raise-cli/src/raise_cli/skills_base/rai-research Command: npx skills add https://github.com/humansys/raise --skill rai-research-humansys
Research decisions often rely on a single source or intuition, risking biased or incomplete conclusions. This Skill enables epistemologically rigorous, triangulated research to inform ADRs and evaluations.
Execute a standard-depth topic prompt to produce a complete evidence catalog, synthesis, and recommendations for your ADR.
High-intent search queries and answers about installing and using this skill.
Triangulated research for ADRs uses structured evidence collection and source evaluation with RaiSE levels to synthesize findings, exposing confidence levels and contradictions to produce actionable governance recommendations.
Multi-source research replaces intuition for high-stakes architecture choices by applying structured evidence cataloging and RaiSE source evaluation, ensuring decisions are justified by transparent triangulation rather than single-source bias.
Evaluate source reliability for technical decision-making research using RaiSE criteria, which assigns confidence levels from Very High to Low to structure your evidence catalog and guide transparent triangulation.
To feed architecture decision records, generate an evidence catalog, a synthesis document detailing contradictions and confidence, and actionable recommendations derived from transparent triangulation.
Relying on a single source risks biased or incomplete conclusions for governance evaluations because it lacks epistemological rigor, whereas triangulated research synthesizes multiple evaluated sources to justify decisions.
Yes, you can execute a standard-depth topic prompt to produce a complete evidence catalog, synthesis document, and actionable recommendations suitable for directly informing your ADRs.
yu-iskwFrames complex problems via ten questions and a scoring-based recommendation.
henrique-simoesConvert research notes and transcripts into structured executive reports.
audunstrandEnforce explicit confidence scoring and evidence audits before presenting conclusions.
not-subhuLaunch parallel subagents for source discovery, triangulation, and credibility scoring.
