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 cyanxxy/nl-tax-agent-skills --skill nl-tax-box3git clone --depth 1 https://github.com/cyanxxy/nl-tax-agent-skillsWrote 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/cyanxxy/nl-tax-agent-skills/nl-tax-box3)<a href="https://agentmods.dev/skills/cyanxxy/nl-tax-agent-skills/nl-tax-box3"><img src="https://agentmods.dev/badge/skills/cyanxxy/nl-tax-agent-skills/nl-tax-box3/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/cyanxxy/nl-tax-agent-skills/nl-tax-box3"><img src="https://agentmods.dev/badge/skills/cyanxxy/nl-tax-agent-skills/nl-tax-box3.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00043 | $0.01548 |
| Opus 5 | $0.00022 | $0.00774 |
| Sonnet 5 | $0.00009 | $0.00310 |
| Haiku 4.5 | $0.00004 | $0.00155 |
Grade A, and why
nl-tax-box3 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 12d ago.
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 — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.
NL Tax Box 3
Background helper for box 3 notes.
Annual 2025 must cover fictitious return and werkelijk-rendement data collection for user review. Provisional 2026 must use only the provisional fictitious method and must never ask for werkelijk rendement.
This helper participates in a conversational workflow. It does not assume all asset and debt inputs are pre-staged. When values are missing, return a structured open-question packet for the calling skill instead of inventing zeros.
This helper may be called through a Skill/Task tool or inlined by an owning workflow when no such tool exists. The same output contract applies either way.
Hard rules
- Annual 2025: collect and compare fictitious return and werkelijk rendement when the user wants the actual-return comparison.
- Ask only branch-applicable annual inputs. A savings-only case needs the 1 January balance for the fictitious method and actual 2025 interest for an actual-return comparison; it does not need a 31 December bank balance merely because the comparison was offered.
- Provisional 2026: use only the fictitious method.
- Never request werkelijk-rendement inputs in a provisional workflow.
- In provisional 2026, accept a debt into
schuldenonly after the official inclusion/exclusion screen. Do not use "all debts except the own-home mortgage" as a shortcut; unresolved debts remain manual-review rows outside accepted totals. - Compute only from values with a real source or an explicitly confirmed assumption.
- The agent classifies each row from the reviewed facts and official rules. Python never infers a category from a description, name, or keyword.
- Before arithmetic, represent every row with
category,status,value, andprovenance. Onlystatus: "accepted"rows inbanktegoeden,overige_bezittingen, orschulden, with finite non-negative values and non-empty provenance, enter trusted totals. Keep every other row in a rejected/manual-review table with a reason. - The bundled scripts require
--partner-full-year-confirmedalongside--has_partnerbefore they double the heffingsvrij vermogen and the schulden drempel;--has_partneron its own raises an error. They also reject negative or non-finite amounts. Do not present a doubled allowance until full-year partnership is confirmed.
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 117 lines · 43 tokens per session scan A 6403a2e75bf6
nl-tax-box3 is a skill published in the GitHub repository cyanxxy/nl-tax-agent-skills (10 stars, last pushed 17d ago), licensed Apache-2.0. It adds 43 tokens to every session and 1,548 once invoked, about $0.0002 per session on Opus 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 2026-08-31.
Other skills, from other repositories
expense-review-policy
Review invoices and contracts against accounts-payable policy before human approval.
kyc-doc-parse
Parse an investor or client onboarding packet into structured KYC fields — identity, ownership, control, source of funds, and document inventory. Use as the first step of KYC screening; output feeds the rules engine.
audit-support
Support SOX 404 compliance with control testing methodology, sample selection, and documentation standards. Use when generating testing workpapers, selecting audit samples, classifying control deficiencies, or preparing for internal or external audits.
fiscaliste
Fiscaliste IA pour la fiscalité personnelle des particuliers français : optimisation et déclaration de l'impôt sur le revenu, IFI, revenus du capital, revenus fonciers, equity salarial, crypto-actifs et PER. Couvre le calcul de l'IR (barème, quotient familial, décote, PAS, CEHR, revenus exceptionnels), la déclaration…
regulatory-analysis
Analyzes documents and processes against FINRA, SEC, Federal Reserve, and CFPB regulatory frameworks. Identifies compliance gaps, classifies findings by severity, and recommends remediation. Use when performing compliance audits, regulatory reviews, gap analyses, or verifying policy adherence to financial regulations.
subcontractor-payment-tracker
Track subcontractor payments, lien waivers, and compliance. Manage payment schedules and documentation.