Route prompts through reasoning lenses and generate compact execution-ready prompts.
npx skills add https://github.com/Ocean326/Agents --skill thinking-lenses
Please help me install this Agent Skill. Skill: thinking-lenses Source: https://github.com/Ocean326/Agents/tree/main/skills/global/thinking-lenses Command: npx skills add https://github.com/Ocean326/Agents --skill thinking-lenses
Optimizes prompts by routing requests into inline-lite, inline-structured, or pre-optimizer modes, then rewrites them using one primary reasoning lens plus one supporting lens to improve clarity, goal alignment, and risk management before execution.
Choose the primary and supporting lenses for the given task and return a compact, execution-ready prompt.
High-intent search queries and answers about installing and using this skill.
To optimize a prompt, you can route it through reasoning lenses that rewrite the request with explicit goals, constraints, and evaluation criteria. This ensures a single primary lens and one supporting lens are selected to produce a compact, execution-ready prompt.
High-risk or high-cost prompts are best handled using a pre-optimizer lens mode. This approach applies tighter constraints and evaluation criteria before execution to improve goal alignment and risk management.
Choose inline-lite for simple prompts to balance speed, and inline-structured for recurring tasks. Lens routing automatically selects the appropriate delivery mode based on task complexity and ambiguity.
Yes, you can apply a dual-lens rewriting strategy by selecting a single primary reasoning lens and one supporting lens. This combination produces a compact, actionable prompt while maintaining clear goal alignment.
Underspecified prompts often yield poor results due to lacking constraints and evaluation criteria. Using trigger-driven workflows with lens-based routing rewrites these prompts to add tighter constraints and explicit goals before execution.
GanyuanRanEnforce test-first routing and verification for software development tasks.
KYRIE66nbClarify vague requests into testable requirements through a Socratic questioning loop.
selcukyucelGenerate structured evaluation datasets with binary rubric scoring for AI outputs.
skkeoriwAutomate AI system development with declarative programming and automated optimization.
