define-referents

Ground ambiguous terms in concrete referents and naming constraints before design or implementation.

Updated May 12, 2026
One-click install
npx skills add https://github.com/mtk177a/skills --skill define-referents-mtk177a
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: define-referents
Source: https://github.com/mtk177a/skills/tree/main/skills/define-referents
Command: npx skills add https://github.com/mtk177a/skills --skill define-referents-mtk177a

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? New or ambiguous terms often collapse distinct concepts—such as a threshold value, a condition, and an event—into one fluent label, silently distorting later design, investigation, or code. This Skill performs a semantic preflight that grounds each material term in a concrete referent before any naming happens. ## Core Features & Use Cases - Grounding-then-Naming workflow: Builds a label-free Grounding table (referent, semantic role, relationships, uncertainty) first, then a Naming table with established terms, proposals, or concrete wording. - Explicit state contract: Reports Ready, Decision required, or Blocked so unresolved public-contract meanings are escalated while low-impact local naming completes without unnecessary confirmation. - Handoff boundary: Returns a referent-and-naming contract to the originating workflow without drafting the downstream document, design, or code, and without creating unauthorized sidecar files. - Use Case: Before naming a public SDK boolean streaming, use this Skill to separate whether it means the caller requested streaming, the server accepted it, bytes are flowing, or streaming completed—then present the decision-relevant question instead of silently choosing. ## Quick Start Before designing this feature, perform a semantic preflight for the term compression point and separate the numeric limit, the reached-limit condition, and the summarization event into a referent table with naming constraints.

Frequently Asked Questions about define-referents

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I define ambiguous terms before naming variables or API fields?▼

Run a semantic preflight that builds a label-free Grounding table first, recording each concrete referent, its semantic role, relationships, and uncertainty. Only after that table validates do you create a Naming table mapping each row to an established term, a proposed term with a first-use definition, or concrete wording.

What is the difference between a Grounding table and a Naming table?▼

A Grounding table describes what each referent actually is—its purpose, role, sequence, and uncertainty—without any candidate labels. A Naming table comes afterward and maps each Grounding ID to a term with a status such as established, proposed, concrete wording, or decision required.

When should I use semantic preflight instead of clarifying the whole request?▼

Use it only when the ambiguity is specifically about referents or terminology that could distort later reasoning, or when a referent table is explicitly requested. Overall request ambiguity about scope, users, or goals belongs to a request-clarification workflow, not this Skill.

Does this Skill write the design document or code after defining terms?▼

No. It returns a referent-and-naming handoff to the originating workflow and stops. It does not draft the target document, design, report, or code, and it never creates a separate referent-table file unless the user explicitly authorizes that extra artifact.

What happens when a term's meaning is undecided or evidence is missing?▼

The Skill reports Decision required when competing meanings affect a public contract or domain boundary, presenting grounded alternatives and the decision question without choosing silently. It reports Blocked when material evidence is unavailable, and never invents a referent, purpose, or naming authority to complete a table.

Does every naming decision require user confirmation?▼

No. When grounding is complete and only a low-impact, local, reversible naming choice remains, the Skill finishes in the same turn with a Ready state. Confirmation is reserved for unresolved semantic choices that change public contracts or domain boundaries.