Structures chat responses with numbered headers, sized output paths, and purpose tracking.
npx skills add https://github.com/alatyshau/duet-work --skill duet-chat-alatyshau
Please help me install this Agent Skill. Skill: duet-chat Source: https://github.com/alatyshau/duet-work/tree/main/skills/duet-chat Command: npx skills add https://github.com/alatyshau/duet-work --skill duet-chat-alatyshau
## Response RX header, making any turn addressable and referenceable later. - Three Output Paths: The assistant chooses between a multi-turn outline plan (O1), a structured sectioned answer over 200 words (O2), or concise prose under 200 words (O3), so output size always matches the task. - Purpose Tracking: A dedicated Purposes section maintains a persistent, ID-tagged list of the chat's goals with status marks (new, achieved, cancelled, split), keeping long discussions honest about what they serve. - Use Case: A user running dozens of parallel advisory chats activates this Skill so each reply is understandable without history, sized to the question, and continuously mapped against the conversation's evolving goals. ## Quick Start Activate the duet-chat skill and conduct this conversation following its numbering, sizing, and purpose-tracking rules.High-intent search queries and answers about installing and using this skill.
Use a response numbering rule where every reply starts with an H2 header like ## Response R1, incrementing per turn. This Skill enforces that format automatically and derives sub-section numbers from the base response number.
Define explicit output paths based on content size: a multi-turn outline plan for very large outputs, numbered markdown sections for answers over 200 words, and plain prose under 200 words with 60-word paragraph caps for short replies.
Purpose tracking maintains a persistent list of the chat's high-level goals, each with a permanent ID like P1 and status marks such as achieved or cancelled. It distinguishes enduring purposes from temporary objectives using survival and concreteness tests.
Yes, by enforcing self-contained text rules: each section's first sentence must stand alone, unfamiliar terms are explained inline, and references to earlier context are restated rather than pointed to. This Skill applies these rules to every reply.
Split a purpose when its parts become distinguishable, such as when a second line of work opens or one part closes while another continues. The parent ID is replaced by children like P1.1 and P1.2, marked with a split indicator.
outfitter-devDefine Outfitter's philosophical voice and values for consistent content creation.
wkentaroRender open task decisions as an ASCII tree with markers and ordered open questions.
mostafa-ismail-2004Organize and synchronize knowledge across memory files and external data stores.
sanity-labsCraft a tailored system prompt that shapes an AI agent's behavior and tone.