Compromise NLP for JavaScript

What it does

When your agent uses it

• Writing or editing JavaScript that imports compromise or calls nlp()
• Extracting people, places, dates or numbers from English text with rules
• Changing verb tense, pluralizing or normalizing text in JavaScript
• Redacting or anonymizing names in a block of text

Example prompts

• “Use compromise to pull all people and organizations out of this article text.”
• “Write a function that converts every sentence in a string to past tense with compromise.”
• “Anonymize the names in ./data/transcripts.txt using compromise.”
• “Build a simple chatbot intent matcher with compromise match patterns.”

Requirements

• The compromise npm package

Workflow steps

5 steps, taken from the step headings in SKILL.md.

1. 1Transforms mutate the document in place — read the result from the original variable
2. 2Only real tags work — an invalid #Tag matches nothing, silently
3. 3Match-syntax is term-level, not regex
4. 4Sentences are the ceiling — matches don't cross sentence boundaries
5. 5compromise is the full build — import it unless size is critical

What it can do on your machine

Read from SKILL.md and the folder at commit 2e7a1b9. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  No scripts in the folder and no shell commands in SKILL.md (its code samples are javascript).

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Compromise NLP for JavaScript loads about 1.8k tokens when it runs. Until then it costs about 189 tokens; SKILL.md has 569 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~189

When it runs · the whole SKILL.md, loaded when a task matches

~1.8k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from spencermountain/compromise at commit 2e7a1b9, republished under its MIT licence (© spencermountain). 569 words, ~1,795 tokens.

name

compromise-nlp

description

Write correct code with the `compromise` JavaScript NLP library (the npm package `compromise`, imported as `nlp`). Use this whenever you are writing or editing JS/TS that imports compromise, calls `nlp(...)`, or chains methods like `.match()`, `.tag()`, `.people()`, `.verbs()`, `.nouns()`, `.numbers()`, `.normalize()`, or `.replace()` — and also whenever doing rule-based natural-language tasks in JavaScript (extracting entities/dates/numbers, matching text patterns, changing verb tense, pluralizing, redacting/anonymizing text, building a simple chatbot intent matcher) where compromise is or could be the tool. compromise's match-syntax and tagset are non-obvious and easy to get wrong from memory; consult this skill before guessing.

import nlp from 'compromise'

let doc = nlp('she sells seashells by the seashore.')
doc.verbs().toPastTense()        // transform
doc.text()                       // 'she sold seashells by the seashore.'

let doc = nlp('I walk to work')
doc.verbs().toPastTense()
doc.text()                                            // ✅ 'I walked to work'  (read from doc)

nlp('I walk to work').verbs().toPastTense().text()    // ❌ 'walked work'  (just the selection)

let past = doc.clone().verbs().toPastTense().text()

nlp('John Smith left').match('[<who>#Person+]').groups('who').text()   // 'john smith'

// entities
nlp(text).people().out('array')        // ['Mary', 'Dr. John Smith']
nlp(text).topics().out('array')        // people + places + organizations
// NOTE: out('array') keeps trailing punctuation ('Paris.'); use .text('normal') for clean tokens

// transform (remember rule #1 — read from doc)
let d = nlp('I walk'); d.verbs().toPastTense(); d.text()      // 'I walked'
let n = nlp('one dog'); n.nouns().toPlural(); n.text()        // 'one dogs'

// numbers
nlp('five hundred').numbers().toNumber().text()   // '500'
nlp('it cost twelve dollars').numbers().get()     // [12]

// find / route (boolean, never mutates)
nlp('the deal is closed').has('#Determiner #Noun')   // true

// teach it words
nlp(text, { kermit: 'FirstName' })                // per-call lexicon
nlp.addWords({ frodo: 'FirstName' })              // global

doc.debug()        // prints how every word was actually tagged — start here
doc.json()         // full structured data: terms, tags, offsets
nlp.verbose(true)  // log the tagger's decision-making

© spencermountain, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 8 other files in docs of spencermountain/compromise.

• SKILL.md
• api.md
• concepts.md
• development.md
• match-syntax.md
• recipes.md
• spec-format.md
• tag-definitions.md
• tagging-differences.md

Open the folder on GitHubat commit 2e7a1b9