# Context Glossary

> Build and maintain a per-repo CONTEXT.md, a short prose dictionary of the project's domain terms the agent reads to decode jargon and name things consistently. Use when project terminology is being discussed or disputed, when naming feels inconsistent, when a hard-to-explain concept keeps needing a paragraph, or when writing or editing a CONTEXT.md or an ADR.

- Skill: `presidenteog/context-glossary` (Agent Skill)
- Install (CLI): `npx skillmds@latest add presidenteog/context-glossary`
- Raw SKILL.md: https://api.skillmd.com/api/skills/presidenteog/context-glossary/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: PresidenteOG (https://skillmd.com/u/presidenteog)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/presidenteog/context-glossary

---


# Context Glossary

Adapted from https://github.com/mattpocock/skills (mattpocock/skills, MIT) — rewritten, not copied.

## Overview
An agent dropped into a repo re-derives the jargon every session and spends twenty words where the team uses one. A `CONTEXT.md` at the repo root fixes this: a short dictionary of the domain's nouns and verbs, each with a one-line definition and the phrases to avoid. Read it at session start. It makes variable, function and file names land on the shared term, makes the codebase easier to navigate, and cuts thinking tokens.

This is the prose layer. `graphify` builds the machine-readable graph of how things relate; `/vault-search` retrieves across the vault. `CONTEXT.md` is the thing a human and the agent both read to agree on what one word means.

## When to use
- A term keeps getting used two ways, or a new synonym enters for something already named.
- Naming in a diff or a plan feels ad hoc.
- Explaining a concept keeps taking a paragraph that a single coined term would replace.
- Writing or revising a `CONTEXT.md` or an ADR.
- Not a substitute for `CLAUDE.md` (rules and workflow) or `graphify` (relationship graph).

## CONTEXT.md shape
| Section | Holds |
|---|---|
| Language | Each term: **bold name**, one-line definition, then `_Avoid_:` with the rejected synonyms |
| Relationships | Plain sentences: "An X holds many Y", "A Y carries one Z at a time" |
| Flagged ambiguities | Terms that meant two things, and how it was resolved |

## The active discipline
- **Challenge every term against the glossary.** New word in conversation: is it a synonym of an existing term (fold it in) or genuinely new (add it, with its avoid-list)?
- **Stress-test with edge-case scenarios.** Walk a weird case aloud. Where the words get clumsy is where a term is missing or wrong.
- **Record hard decisions as ADRs.** A choice that was hard to reach and would look arbitrary later gets a short ADR (context, decision, consequences) next to `CONTEXT.md`.
- **Update inline.** Do not batch glossary edits to the end of the session. Fix the term the moment it shifts.

## Where it lives
Repo root `CONTEXT.md` for a code project under `ClaudeFX/Projects/<name>/`. For vault-level domain work the vault's own conventions in `ClaudeFX/CLAUDE.md` win.