Skip to content
Agentmart
Submit a tool

ubiquitous-language

Use when crystallizing a domain glossary for a single bounded context — extract canonical terms from replay JSONL, evidence JSONL, and src/ identifiers, then…

Skill by mattmre

IndexedMIT★ 3
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/mattmre-ubiquitous-language

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Use when crystallizing a domain glossary for a single bounded context — extract canonical terms from replay JSONL, evidence JSONL, and src/ identifiers, then write GLOSSARY.md scoped to that context only.

Use cases: writing

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/ubiquitous-language

What's inside

Skill name
ubiquitous-language
Description
Use when crystallizing a domain glossary for a single bounded context — extract canonical terms from replay JSONL, evidence JSONL, and src/ identifiers, then write GLOSSARY.md scoped to that context only.
License
MIT
Instructions
247 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Ubiquitous Language (Per Bounded Context)

## When to use this skill

Use this skill when a single named bounded context (from
`docs/adr/0005-bounded-contexts.md`) needs a crisp glossary — typically
right after `zoom-out` reveals two callers using the same word for
different concepts, or as part of a domain-modelling pass before refactor.
The trigger is "we need a canonical glossary for context X". This skill
REFUSES to run without an explicit context name; there is no repo-wide
glossary mode.

## Bounded Context Required

This skill is hard-coupled to the EVOKORE-MCP context map in
[`docs/adr/0005-bounded-contexts.md`](../../../docs/adr/0005-bounded-contexts.md).
The eight bounded contexts are:

1. **Skill Registry & Discovery** — `src/SkillManager.ts`,
   `src/RegistryManager.ts`, `src/ToolCatalogIndex.ts`,
   `src/ToolCatalogPagination.ts`, `src/ProfileResolver.ts`,
   `src/rerank/successRerank.ts`.
2. **Proxy & Routing** — `src/ProxyManager.ts`,
   `src/utils/resolveCommandForPlatform.ts`, `src/httpUtils.ts`.
3. **Auth & Security** — `src/auth/OAuthProvider.ts`,
   `src/SecurityManager.ts`, `src/ComplianceChecker.ts`,
   `src/ContainerSandbox.ts`.
4. **Session & Continuity** — `src/SessionIsolation.ts`,
   `src/SessionManifest.ts`, `src/SessionManifest.schema.ts`,
   `src/SessionStore.ts`, `src/stores/MemorySessionStore.ts`,
   `src/stores/FileSessionStore.ts`, `src/stores/RedisSessionStore.ts`,
   `src/HttpServer.ts`.
5. **Orchestration & Fleet** — `src/OrchestrationRuntime.ts`,
   `src/FleetManager.ts`, `src/ClaimsManager.ts`,
   `src/WorkerManager.ts`, `src/WorkerScheduler.ts`,
   `src/TrustLedger.ts`.
6. **Audit & Webhooks** — `src/AuditLog.ts`, `src/AuditExporter.ts`,
   `src/WebhookManager.ts`, `src/PluginManager.ts`.
7. **Telemetry & Analytics** — `src/TelemetryManager.ts`,
   `src/TelemetryExporter.ts`, `src/TelemetryIndex.ts`,
   `src/SessionAnalyticsManager.ts`, `src/NavigationAnchorManager.ts`.
8. **Memory & Knowledge** — `src/MemoryManager.ts`.

The voice subsystem is NOT a bounded context (per ADR-0005's
"Out-of-context" section); this skill rejects `--context voice`.

The context name (slug form, kebab-case) MUST be passed as `--context
<slug>` (or equivalent argv) when invoking. Accepted slugs are:
`skill-registry-and-discovery`, `proxy-and-routing`,
`auth-and-security`, `session-and-continuity`,
`orchestration-and-fleet`, `audit-and-webhooks`,
`telemetry-and-analytics`, `memory-and-knowledge`.

If no `--context` is provided, this skill MUST refuse to run, emit an
error pointing the caller at `docs/adr/0005-bounded-contexts.md`, and
exit without writing any output. There is NO fallback to a repo-wide
`UBIQUITOUS_LANGUAGE.md` — that file is an explicit anti-pattern in
EVOKORE-MCP.

## Adapted From Upstream

This adapter is derived from
[`ubiquitous-language/SKILL.md`](../../upstream/mattpocock-skills/ubiquitous-language/SKILL.md)
in the `mattpocock/skills` submodule pinned at
`90ea8eec03d4ae8f43427aaf6fe4722653561a42`. The upstream extracts a
DDD-style glossary from "the current conversation" and writes one
file at the project root.

To read the upstream body without loading the full file, prefer:

```text
nav_read_anchor SKILLS/upstream/mattpocock-skills/ubiquitous-language/SKILL.md <anchor-id>
```

License: MIT, Copyright (c) 2026 Matt Pocock. See repo-root `NOTICE` and
`SKILLS/upstream/mattpocock-skills/LICENSE`.

## EVOKORE-Specific Adaptations

Concrete deltas vs the upstream skill:

- **No repo-wide glossary.** The upstream writes a single
  `UBIQUITOUS_LANGUAGE.md` at the project root. EVOKORE-MCP rejects
  this as a DDD anti-pattern (see ADR-0005 "Decision" section: the
  same word legitimately means different things across contexts). This
  adapter writes one glossary per context only.
- **No interactive interview.** The upstream "scans the conversation"
  and proposes terms in a chat dialog. EVOKORE is autonomous — this
  adapter reads append-only artifacts instead: replay JSONL, evidence
  JSONL, and `src/` identifiers filtered to the named context's
  modules.
- **Always context-scoped.** The first line of every glossary file is
  the bounded-context name and a back-link to ADR-0005. There is no
  unscoped run mode.
- **Source citations are required.** Each term MUST cite at least one
  source: a `src/<file>.ts` identifier, a replay JSONL line range, or
  an evidence JSONL line. The upstream allows undocumented opinion;
  EVOKORE requires evidence (per the project's "Evidence Capture" hook
  philosophy).
- **Composition phrasing.** When a glossary surface ambiguity that
  warrants an architectural change, the skill emits literal "invoke
  improve-codebase-architecture skill" phrasing so the static
  composition graph (`derive-skill-composition.js`) can edge to it.
- **Trigger-explicit description.** Frontmatter `description` is
  rewritten in EVOKORE's "Use when ..." form so `resolve_workflow`
  semantic ranking can rank this skill above deep reference leaves.

## Output Path

Glossary files are written under this skill's own directory so they
live alongside the rules that produced them and cannot be confused
with a repo-wide artifact:

```text
SKILLS/CONTEXT/ubiquitous-language/glossaries/<context-slug>.md
```

Where `<context-slug>` is the kebab-case slug listed in the **Bounded
Context Required** section above. Existing files are read first and

From mattmre/EVOKORE-MCP-PUBLIC, MIT.

Signals

Stars
3
Forks
0
Open issues
0
Last commit (this folder)
May 21, 2026 (5 months ago)
Last release
May 22, 2026 (5 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from mattmre

Cite this page

ubiquitous-language skill. Agent Mart. https://agentmart.dev/skills/mattmre-ubiquitous-language (last updated 2026-10-24)