Transform vague targets into concrete seeds and strategic plans.
npx skills add https://github.com/cyberAlchemyAI/Arcanum --skill refine-cyberalchemyai
Please help me install this Agent Skill. Skill: refine Source: https://github.com/cyberAlchemyAI/Arcanum/tree/main/.claude/skills/refine Command: npx skills add https://github.com/cyberAlchemyAI/Arcanum --skill refine-cyberalchemyai
Refine provides a governed discovery and design orchestration workflow that turns vague targets into concrete seeds, designs, or non-executed plans through the canonical Arcanum refinement loop.
Propose a Refine Run Strategy and wait for explicit confirmation before executing the canonical loop.
High-intent search queries and answers about installing and using this skill.
To transform vague targets into concrete plans, a governed refinement loop generates strategic seeds through auditable dispatch routes. This ensures runtime handoff and stage evidence indexing occur before execution begins.
A refinement loop is a canonical ten-stage orchestration process supporting discovery, critique, repair, and design scenarios. It applies dispatch-spec validation and native-receipt-based stage provenance to produce refined synthesis.
To start a refinement run, propose a Refine Run Strategy and wait for explicit confirmation. This triggers canonical loop orchestration to process your critique or repair targets with proper governance boundaries.
No, the refinement workflow requires no external dependencies. It operates independently using internal scripts to orchestrate discovery, dispatch-spec validation, and final synthesis generation.
The best way to maintain auditable stage provenance is using native-receipt-based tracking within a canonical ten-stage refine loop. This automatically indexes stage evidence and validates dispatch routes throughout the design process.
You should not use a governed refinement loop when your task requires immediate execution without strategic planning, runtime handoff, or ownership boundaries. It is designed for non-executed plan generation and risk reduction, not direct deployment.
Jawad-ChaudharyRead a current spec, apply feedback, and generate an updated spec.
navraj007inEvaluate architecture quality across six pillars and generate remediation plans.
GrinRusDefine stage-specific research contracts for subagent evidence reading and handoff.
AlexmacappleTransforms vague words into multilingual semantic seeds for AI prompts.
