Compromise is a lightweight natural-language processing library that turns text into data by tokenizing words and sentences, identifying basic grammar, and exposing structured document operations. It is for JavaScript applications that need simple text parsing, tagging, and extraction in the browser or on a server.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add spencermountain/compromise --skill pos-tagging-fixesgit clone --depth 1 https://github.com/spencermountain/compromiseWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/spencermountain/compromise/pos-tagging-fixes)<a href="https://agentmods.dev/skills/spencermountain/compromise/pos-tagging-fixes"><img src="https://agentmods.dev/badge/skills/spencermountain/compromise/pos-tagging-fixes/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/spencermountain/compromise/pos-tagging-fixes"><img src="https://agentmods.dev/badge/skills/spencermountain/compromise/pos-tagging-fixes.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00072 | $0.04159 |
| Opus 5.5 | $0.00029 | $0.01664 |
| Sonnet 5.5 | $0.00014 | $0.00832 |
| Haiku 4.5 | $0.00007 | $0.00416 |
Grade A, and why
pos-tagging-fixes scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 291 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Internal POS-tagging fixes
Work from the repository root. Read its AGENTS.md and preserve concurrent changes. This skill does not authorize dependency changes, existing-test edits, commits, or PRs. If generating tests, write cases and expectations blindly from the requested language behaviour before researching the implementation; do not derive expected results from the current tagger.
Choose the smallest intervention that expresses the linguistic distinction at the stage where its evidence exists. Runtime, filesize, and behavioural scope are separate costs. A one-line switch assignment can change more behaviour than a longer contextual rule. General postTagger rules are by far the slowest intervention; sometimes their wider context is necessary. Hand-made preTagger loops can be fast but add substantial code for niche cases. Prefer an existing data table or indexed rule when it expresses the same condition faithfully.
Establish the failure
Record the input, target term, expected tags, and contrasting readings that must remain valid.
Use compromise's conventions in tagging differences and
tag definitions, not a different tagger's labels. Distinguish a
wrong POS from a tokenization, contraction, entity-selection, or chunking failure. Correct tags
with a wrong .verbs() selection may require a selection fix rather than another tagging rule.
Reproduce with ESM imports from ./src/two.js or ./src/three.js. Avoid stale CommonJS builds.
Start with a word-filtered trace, enabling it before parsing:
import nlp from './src/three.js'
nlp.verbose('tagger', { word: 'walk' })
const doc = nlp('They walk home.')
nlp.verbose(false)
doc.debug()
console.dir(doc.json(), { depth: null })
Trace both the first wrong decision and the last rule to touch the term. Search verbose reason
strings with rg in src/. Tagger events report actual added and removed tags, including
conflict removals and implied parents. They omit no-ops and blocked requests: absence of an
event does not prove that a rule never ran. Check frozen state and conflicts separately.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today First seen · 291 lines · 72 tokens per session scan A c95b7d53270a
pos-tagging-fixes is a skill published in the GitHub repository spencermountain/compromise (12,164 stars, last pushed yesterday), licensed MIT. It adds 72 tokens to every session and 4,159 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-19.
Other skills, from other repositories
spacy-nlp-pipeline
Use this Skill to build NLP pipelines with spaCy: tokenization, POS tagging, dependency parsing, NER, custom rule components, and multi-language model selection.
trace
Evidence-driven tracing lane that orchestrates competing tracer hypotheses in Claude built-in team mode.
smiles-validation
Strict SMILES validation, structural comparison, and modification verification. Catches invalid LLM-generated molecules.
nonlinear-solvers
Select and configure nonlinear solvers for f(x)=0 or min F(x). Use for Newton methods, quasi-Newton (BFGS, L-BFGS), Broyden, Anderson acceleration, diagnosing convergence issues, choosing line search vs trust region, and analyzing Jacobian quality.
linear-solvers
Select and configure linear solvers for systems Ax=b in dense and sparse problems. Use when choosing direct vs iterative methods, diagnosing convergence issues, estimating conditioning, selecting preconditioners, or debugging stagnation in GMRES/CG/BiCGSTAB.
numerical-stability
Analyze and enforce numerical stability for time-dependent PDE simulations. Use when selecting time steps, choosing explicit/implicit schemes, diagnosing numerical blow-up, checking CFL/Fourier criteria, von Neumann analysis, matrix conditioning, or detecting stiffness in advection/diffusion/reaction problems.