Vocabulary logo

Vocabulary

Community
xiaolai
vocabulary

Use when writing, reviewing, or naming any cc-suite artifact — pick the canonical noun or verb from this registry rather than coining a synonym. Loaded by NLPM's scorer and checker when R51 is enabled in .claude/nlpm.local.md.

Overview

Publisherxiaolai
Repositorycc-suite
Skill namevocabulary
Stars
56
Forks
21
Bundled files
1
LicenseISC
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by xiaolai on GitHub. Read the source before you install it.

Installation

Install the Vocabulary AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/xiaolai/cc-suite/tree/main/skills/cc-suite/vocabulary
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/xiaolai/cc-suite.git /tmp/cc-suite
mkdir -p .claude/skills
cp -r /tmp/cc-suite/skills/cc-suite/vocabulary .claude/skills/vocabulary
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Vocabulary in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Vocabulary on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Vocabulary is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

cc-suite Domain Vocabulary

The canonical noun-and-verb set for cc-suite. Every artifact name, prose description, and rule wording should draw from this registry. Drift between canonical terms and declared deprecated synonyms is flagged by /nlpm:check and penalized by /nlpm:score when R51 is enabled.

Status: seeded by /nlpm:vocab-init on 2026-05-26 and pruned during the initial migration pass. Only deprecation pairs that can be cleanly enforced (no false positives in the current corpus) are in registry.yaml. Other observed drift is documented below as convention-only with the rationale for non-enforcement.


Scopes

cc-suite has a single scope. Unlike NLPM (internal + auditor), cc-suite's artifacts all describe one system — the bridge between Claude Code, Codex CLI, and Antigravity CLI (agy).

ScopePathsDescription
internalcommands/, skills/cc-suite/, templates/agents/, AGENTS.mdThe cc-suite bridge subsystem itself.

Enforced deprecation pairs (R51 will flag these)

These are clean drifts: the canonical term has fully replaced the deprecated one in the corpus, and the deprecated term has no remaining legitimate uses.

CanonicalDeprecated synonymStatus
dimensionpillarFully migrated (Mar 2026). All 5 audit-family commands now use Dimension N: headings and (N dimensions) phrasing.
subagentsub-agentFully migrated. The hyphenated form was a typo accident in agent-design SKILL.md and audit-agent.md.
refreshfreshenFully migrated. freshen-aware / freshen semantics in AGENTS.md rewritten to refresh-aware / refresh semantics.

Canonical-only verbs (no deprecation; just the canonical form)

Used to establish the cc-suite verb vocabulary so new artifacts pick from this set rather than coining synonyms. R51 doesn't penalize anything here because no observed drift exists.

CanonicalDefinition
auditStructured multi-dimension code/artifact scan with severity table (Codex-delegated).
bridgeUmbrella verb for making one tool's config visible to another.
updateUser-side coupled re-render after claude plugin update cc-suite (slash command).
initFirst-time project setup of all cc-suite bridge artifacts.
checkLightweight prerequisite / connectivity probe.
consultInvoke a cc-suite advisor and receive its judgment.
delegateForward a task to Claude Code via mcp__claude-code__claude_code.
fixCode-level remediation of a specific reported defect (distinct from repair).
repairNon-interactive re-run of all bridge / registration scripts (distinct from fix).
verifyConfirm a prior code-fix outcome via Claude MCP.
registerPersist an MCP server entry into .mcp.json or .codex/config.toml.
exposeSymlink-based sub-operation of bridge for cc-suite plugin skills.
mirrorOne-directional copy sub-operation of bridge for hooks and MCP.
setupHigher-level user activity composed of init + configure + verify.
reviewSend a plan or PR to Codex for architectural review (distinct from consult).
cancel continue implement plan analyze result status add remove list preflight unbridgeSingle-command verbs; one canonical, no competing synonyms observed.

Canonical-only nouns

CanonicalDefinition
advisorA project-scoped value-over-rules persona declared in .cc-suite/agents/, backed by claude-octopus as a separately-configured MCP server.
bridgeThe cc-suite-managed subsystem that wires Claude Code, Codex, and Antigravity (agy) together.
findingA single defect surfaced by an audit/review — file:line, severity, recommendation. Use in structured audit output (column headers, "Top Findings" sections, per-finding iteration in audit-fix/verify).
sentinel blockThe cc-suite-owned section in .codex/config.toml or .gitignore, delimited by sentinel comments.
threadThe Codex execution-continuation handle (the threadId MCP parameter).
registrationThe persisted MCP server entry produced by bridge.
presetA starter advisor agent file shipped under templates/agents/.
pinThe version string for claude-octopus in scripts/lib/claude-octopus-pin.txt.
timelineThe per-advisor persistent consultation history at .cc-suite/agents/<name>/timeline/.
artifactA file cc-suite owns — command, skill, agent, hook, manifest, config.
scriptAn executable under scripts/ — bash, python, or mjs.

Role-noun suffixes for advisor preset names (intentional sub-variants)

These are not drift — each suffix encodes the persona's stance.

SuffixStanceExample preset
_advisorGeneral-purpose value advisor; broad-scope judgment.north_star_advisor
_reviewerReviews finished work against a single quality dimension.clarity_reviewer
_criticJudges the output of an upstream process (typically docs).documentation_critic
_advocateArgues for one specific action over alternatives.deletion_advocate, simplicity_advocate
_skepticAdversarial reviewer; assumes hostile inputs / worst case.security_skeptic

When introducing a new preset, pick the suffix that best matches its stance. Default to _advisor when none fit.

Why these pairs are convention-only (not enforced by R51)

The /nlpm:vocab-drift scan surfaced the drift candidates listed below. Each looks like a clean canonical/deprecated pair but turns out to be polysemous in cc-suite. R51 is mechanical — it flags every occurrence of a deprecated term regardless of context — so declaring these would generate false-positive penalties on legitimate uses. They are conventions, not rules:

Concept pairWhy we don't enforce
finding ← issuefinding is canonical for audit-output rows (column headers, Top Findings sections, per-finding iteration in audit-fix.md / verify.md). But issue legitimately means a different thing elsewhere: infrastructure problem in diagnose.md ("Found N issues. Fix them?"), implementation hiccup in implement.md ("Issues encountered"), code-quality category in audit prose ("minor clarity issues"). A blanket deprecation would mis-rename all three contexts. Convention: structured rows say Finding; everything else stays issue.
advisor ← agentadvisor is canonical for cc-suite personas in prose. But agent is unavoidable in file paths (.cc-suite/agents/, templates/agents/), command names (add-agent, remove-agent, list-agents) — chosen to match Claude Code's .claude/agents/ convention — and refers to Claude Code's native subagent concept in design discussions. Convention: speak about "advisors" in prose; keep the file/dir/command layout using agent.
thread ← sessionthread is canonical for the Codex execution-continuation handle. session appears in continue.md's description as user-facing language ("Continue a previous Codex session"). Convention: prefer thread in technical contexts; tolerate session in user-facing descriptions.
delegate ← ask / senddelegate is canonical in skill bodies. ask and send appear in description: frontmatter openings of the claude-* skills because they're more natural English for a one-line summary. Convention: skill body uses delegate; frontmatter description allows either.
preset ← templatepreset is the user-facing concept. template is the directory name (templates/agents/) and appears in add-agent.md frontmatter where it accurately describes what gets copied. Convention: speak about "presets" in prose; the on-disk word is template.
verify ← checkverify is canonical for code-fix-outcome confirmation. check covers infrastructure probes (Step 2: Run additional deep checks in diagnose.md). Convention: don't conflate.

How to extend

  1. Re-run extraction:

    bash
    python3 ${CLAUDE_PLUGIN_ROOT:-~/.claude/plugins/cache/xiaolai/nlpm/1.0.0}/analysis/scripts/extract-vocabulary.py \
      --root . \
      --scopes analysis/vocab-init-scopes.json \
      --out analysis/vocabulary-extract/
  2. New terms appear in analysis/vocabulary-extract/summary.md.

  3. Add a row to the matching table above (Verbs or Nouns, in the correct scope). Cite at least one file as evidence.

  4. To deprecate a synonym, add it to the canonical term's deprecated: list in registry.yaml AND update the "Enforced deprecation pairs" table above. Before doing this, run grep -r "<deprecated>" and confirm the term has no legitimate alternate meanings — if it does, document the pair in "Why these pairs are convention-only" instead.

Adopting R51

To turn on vocabulary drift detection on this project, add to .claude/nlpm.local.md:

yaml
rule_overrides:
  R51:
    enabled: true
    vocabulary_skill: skills/cc-suite/vocabulary/

The pruned registry should produce zero R51 findings on the current corpus.

See also

  • The corpus extractor: ${NLPM_ROOT}/analysis/scripts/extract-vocabulary.py
  • The six design principles: ${NLPM_ROOT}/analysis/vocabulary-design-principles.md
  • Run /nlpm:vocab-drift periodically to surface new candidate pairs as the corpus grows.

Example Invocations

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Vocabulary AI skill do?

Use when writing, reviewing, or naming any cc-suite artifact — pick the canonical noun or verb from this registry rather than coining a synonym. Loaded by NLPM's scorer and checker when R51 is enabled in .claude/nlpm.local.md.

Why use Vocabulary on TypingMind?

Because you install it once and use it with any model. Vocabulary is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Vocabulary in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/xiaolai/cc-suite/tree/main/skills/cc-suite/vocabulary. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Vocabulary?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Vocabulary?

As many as you like. As long as a model supports skills, you can use Vocabulary with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Vocabulary AI skill free?

Yes. It is published on GitHub by xiaolai under the ISC license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇