Aggregate upstream analysis claims into a cited narrative with consensus notes.
npx skills add https://github.com/hellonish/singularity --skill synthesis-hellonish
Please help me install this Agent Skill. Skill: synthesis Source: https://github.com/hellonish/singularity/tree/main/SKILLS/tier2_analysis/synthesis Command: npx skills add https://github.com/hellonish/singularity --skill synthesis-hellonish
Synthesis addresses the need to fuse disparate information streams into a single, coherent narrative. It preserves source citations and avoids cherry-picking, enabling traceable, aggregated insight from multiple upstream analyses.
Provide at least two upstream analysis outputs containing claims with citations to be synthesized into a unified analysis.
High-intent search queries and answers about installing and using this skill.
To synthesize findings from multiple upstream analyses, provide at least two upstream analysis outputs. The Skill aggregates claims, clusters them by theme, and produces a unified narrative with traceable citations and consensus notes.
The best way to aggregate evidence and identify consensus is by clustering upstream claims by theme. This process highlights contradictions and complementary details, presenting a synthesized narrative with explicit source citations.
You combine multiple analysis outputs without cherry-picking by using a synthesis process that preserves source citations. It fuses disparate information streams into a coherent narrative while maintaining traceable, aggregated insight.
Evidence synthesis works for risk assessment and decision-support scenarios. It applies across domains requiring aggregated evidence, turning diverse findings into a cohesive narrative with hedging and clearly linked claims.
To maintain citation traceability when merging claims, the synthesis process produces an AnalysisOutput with explicitly linked claims and source citations. This avoids cherry-picking and ensures aggregated insight remains traceable.
UfukNodeRewrite downstream narrative beats to maintain continuity after an upstream insert.
kju4qSurfaces hidden assumptions and evaluates multiple perspectives for high-stakes decisions.
PostHogCluster LLM traces and generations to identify usage patterns.
mattmacleod16-svgCombine and deduplicate results from diverse sources into a single answer with citations.
