---
name: Arbitration Clause Design And Review
slug: arbitration-clause-design-and-review
category: Writing
description: Arbitration Clause Design And Review helps draft, review, and stress-test commercial arbitration clauses. Use it to spot issues, propose fixes, and assess seats, rules, institutions, and multi-party structures.
github: "https://github.com/lawve-ai/awesome-legal-skills/tree/main/skills/arbitration-clause-design-and-review-hafez-virjee"
language: Python
stars: 662
forks: 83
install: "npx degit https://github.com/lawve-ai/awesome-legal-skills/tree/main/skills/arbitration-clause-design-and-review-hafez-virjee ~/.claude/skills/arbitration-clause-design-and-review-hafez-virjee"
installs_to: ~/.claude/skills/arbitration-clause-design-and-review-hafez-virjee
source_path: skills/arbitration-clause-design-and-review-hafez-virjee/SKILL.md
collection_size: 25
category_size: 1012
collection_url: "https://dirskills.com/collections/lawve-ai/awesome-legal-skills"
added: 2026-08-24T05:17:10.013Z
last_synced: 2026-08-24T05:17:10.013Z
canonical_url: "https://dirskills.com/skills/arbitration-clause-design-and-review"
---

# Arbitration Clause Design And Review

Arbitration Clause Design And Review helps draft, review, and stress-test commercial arbitration clauses. Use it to spot issues, propose fixes, and assess seats, rules, institutions, and multi-party structures.

**Install:**

```bash
npx degit https://github.com/lawve-ai/awesome-legal-skills/tree/main/skills/arbitration-clause-design-and-review-hafez-virjee ~/.claude/skills/arbitration-clause-design-and-review-hafez-virjee
```

## README

# Arbitration Clause Design and Review
## Purpose
Use this skill to help users draft, review and stress-test arbitration clauses in commercial contracts.
The skill is designed to produce practical, usable outputs:
- a clean arbitration clause;
- a concise review of an existing clause;
- a severity-rated list of issues;
- proposed fixes or revised wording;
- optional reasoning, recommendation reports and negotiation arguments.
The skill should be quick and frictionless for non-specialist users, while allowing sophisticated users to request deeper analysis.
## Subtitle
A commercial arbitration workflow for drafting, reviewing and stress-testing dispute resolution clauses.
## Author
Hafez Virjee
## Methodological note
This workflow draws on practical arbitration experience, arbitral-procedure design, and the Delos GAP's focus on arbitral seats, enforcement, legal specificities and cross-border dispute planning.
It is a drafting and issue-spotting workflow. It is not legal advice. Arbitration clauses can have significant consequences under the law of the seat, the governing law, and the laws of enforcement jurisdictions. The user should obtain legal advice before finalising the clause, especially where the transaction is high-value, complex, cross-border, regulated, or involves state-linked, sovereign, mandatory-law or enforcement-sensitive issues.
# When to use this skill
Use this skill when the user asks to:
- draft an arbitration clause;
- review an arbitration clause;
- improve a dispute resolution clause;
- identify pathologies in an arbitration agreement;
- assess whether an arbitration clause is workable;
- select or assess a seat of arbitration;
- select or assess arbitral rules or an arbitral institution;
- produce a clause for a commercial contract;
- generate internal or counterparty-facing arguments for an arbitration clause;
- stress-test whether a clause fits the commercial context.
This skill is for commercial arbitration clauses in contracts.
# When not to use this skill, or when to refer out
Do not attempt to provide a complete solution where the matter primarily involves:
- investment arbitration or treaty-based dispute resolution;
- consumer arbitration;
- employment arbitration where mandatory law may restrict arbitration;
- sports, disciplinary or regulatory arbitration;
- sanctions-heavy or export-control-sensitive matters;
- complex multi-contract or project-finance structures requiring bespoke consolidation or joinder analysis. Note: the skill may still provide the general commercial arbitration architecture for multi-party scenarios; the exclusion applies to bespoke consolidation or joinder analysis, not to the architecture itself.
Where one of these issues arises, do not simply refuse to help unless the whole task is outside scope. Instead:
1. assist with the commercial arbitration clause to the extent appropriate;
2. identify the specific issue that requires specialist advice;
3. explain why that issue matters;
4. recommend that the user obtain legal advice before finalising the clause.

## Multi-party and multi-contract scenarios
Where a scenario involves multiple parties, multiple contracts, or both, provide the commercial arbitration architecture first before flagging specialist referral. Do not retreat to a specialist referral without first providing the framework.

The architecture guidance for complex multi-party scenarios should include, where relevant:
- institution selection: apply the general institution-selection criteria, with the additional factor that the chosen institution's rules on consolidation and joinder should be assessed for compatibility with the multi-party structure. Several major institutions have well-developed multi-party frameworks; the right choice depends on the parties, geography, value, and priorities as in any other scenario;
- a principal arbitration clause in the main contract, with back-to-back arbitration clauses in related contracts that mirror the seat, institution, and rules;
- consolidation and joinder provisions, where the chosen institution's rules support them;
- whether the scenario raises issues outside scope (bespoke project finance structure, non-identical parties across multiple tiers) that require specialist advice.

Flag specialist referral after providing the framework, not instead of it. Do not produce detailed consolidation or joinder wording unless the user requests it or the structure is straightforward enough to make standard wording appropriate.
# State-linked, SOE and public-counterparty contracts
Contracts involving SOEs, state entities, public bodies or sovereign-related parties are not automatically outside scope. Many such contracts are ordinary commercial contracts, and the skill may assist with arbitration clause design where the party appears to be acting in a commercial capacity.
However, if immunity, authority, capacity, procurement law, public law, sanctions, enforcement, treaty protection or other sovereign-related issues may be relevant, the skill should flag those issues specifically and recommend specialist legal advice before finalising the clause.
Do not refuse to assist simply because a counterparty is an SOE or public body. Instead:
1. assist with the commercial arbitration clause to the extent appropriate;
2. identify the specific state-linked issue that requires legal review;
3. explain briefly why that issue matters;
4. avoid giving a definitive view on immunity, authority, public-law validity or enforcement unless properly supported.
Example wording:
I can help with the commercial arbitration architecture, but because the counterparty is state-linked, the clause should be reviewed for authority, capacity, immunity, enforcement and any applicable procurement or public-law requirements before it is finalised.
If the contract is a concession, public-private partnership, treaty-linked arrangement, regulated infrastructure project, or another arrangement where the state is acting in a sovereign or public-law capacity, treat the matter as higher risk. The skill may still identify drafting components or provide a provisional clause, but it must clearly state that specialist advice is required before use.
# Core design principle
Follow progressive disclosure.
## Effort mode and skill instructions
This skill contains specific analytical instructions that override default model behaviour and weightings. These instructions apply regardless of the complexity of the request, the brevity of the user's prompt, or any inference that a "lighter" response is appropriate.

In particular:
- The governing law analysis (see below) must always follow the structured framework set out in this skill, not default model assumptions about neutral or internationally recognised governing laws.
- The institution selection and cost sense-check instructions (see below) must always be followed as analytical steps, including the internal Delos cost sense-check, regardless of how simple or routine the request appears. Whether the result is surfaced publicly depends on the public-surfacing criteria in the institution selection section.

Do not treat these as optional steps that can be skipped when the user's request is short or the context seems straightforward. The governing law framework and cost sense-check are mandatory analytical steps. The cost comparison should be surfaced in the output only where the public-surfacing criteria in the institution selection section are met.

Default behaviour:
1. Ask as few questions as possible.
2. Use uploaded or pasted documents first.
3. Extract relevant context before asking follow-up questions. Before deciding what information is missing, extract from the prompt or document any facts that bear on claimant/respondent posture, relationship duration, payment structure, governing law, seat, institution and likely enforcement. A question is warranted only where the missing information would materially change the clause architecture and cannot be inferred from what is available.
4. Ask only for information that materially affects the clause.
5. Give the practical answer first.
6. Offer deeper reasoning only if requested.

For software distribution, licence, and other recurring-revenue contracts, payment structure (event-driven, periodic, milestone-based) is a key indicator of likely claimant/respondent posture. Extract this from available context where possible. Ask a targeted question only if it cannot be inferred.

The user should feel that the workflow knows what it is doing. Do not overwhelm the user with arbitration theory or institutional detail unless they ask for it.
# Initial intake
Begin by determining the user's objective.
Ask:
Are you looking to draft a new arbitration clause, or review an existing clause?
Then identify the user's role, using this order:
1. in-house counsel;
2. senior lawyer / arbitration practitioner;
3. junior lawyer / trainee;
4. business user / commercial lead;
5. other.
The user's role affects tone and level of explanation, not the quality of analysis.
Invite the user to upload or paste any available material, such as:
- the draft clause;
- the relevant contract;
- a term sheet;
- a deal summary;
- negotiation comments;
- the counterparty's proposed wording;
- user instructions.
Use a document-first approach. Read what is available, extract context, and ask only for missing information that materially affects the analysis.
# Core information to extract
Where available, extract or ask for:
- contract type;
- industry or sector;
- parties and their jurisdictions;
- places of performance;
- governing law of the contract;
- proposed seat of arbitration;
- proposed institution and rules;
- contract value;
- expected dispute value, or an illustrative dispute value;
- likely enforcement jurisdictions;
- whether the relationship is one-off, repeat, long-term or ongoing;
- whether preserving the relationship matters;
- whether the negotiation has been cooperative or acrimonious;
- whether one party is more sophisticated or better resourced;
- whether urgent temporary relief may be needed;
- whether urgent final determination may be needed;
- whether confidentiality is desired;
- whether there are related contracts with non-identical parties;
- whether the user or their client is more likely to be claimant, respondent, or either.
Do not ask all of these questions automatically. Ask only what is missing and material.
# Commercial posture assessment
Assess the commercial function of the arbitration clause.
Use a user-friendly question such as:
In this transaction, if something goes wrong, who is more likely to need to bring a claim - you, the other side, or is it genuinely hard to say?
Use the answer to understand whether the clause should prioritise:
- speed;
- cost predictability;
- access to justice;
- procedural neutrality;
- relationship preservation;
- recognition and institutional familiarity;
- a final answer quickly;
- a more heavyweight process for high-value or complex disputes.
Remain neutral. Do not encourage abusive or bad-faith drafting. It is acceptable to recognise that different commercial postures legitimately affect the appropriate dispute resolution mechanism.
# Main paths
There are two main paths:
1. **Design Path** - drafting a new arbitration clause.
2. **Review Path** - reviewing, stress-testing or improving an existing arbitration clause.
# Design Path
Use the Design Path when the user wants a new clause.
## Design Path default output
The default output should be:
1. clean draft arbitration clause;
2. one-line explanation of the key choices;
3. confidence and missing information box;
4. offer to generate a fuller recommendation report.
In Design Path outputs, the draft clause appears first. Do not place reasoning, analysis, or background before the clause. A user who wants only the clause should be able to read it immediately. Analysis and reasoning follow, kept to the minimum needed to explain the key choices. In Review Path outputs, the overall assessment appears first. The clause should be as short as possible while remaining complete. Do not restate matters already covered by the selected institutional rules unless there is a specific reason to do so.
## Drafting hierarchy
When an institution is selected, use that institution's own model clause as the starting point where available.
Use this hierarchy:
1. If institutional arbitration is selected, use the selected institution's recommended model clause as the base.
2. If Delos arbitration is selected, use the Delos model clause as the base.
3. If UNCITRAL ad hoc arbitration is selected, use the UNCITRAL model clause as the base.
4. If pure ad hoc arbitration is selected, draft from first principles and flag that specialist legal advice may be appropriate.
5. If confidentiality is desired, include an express confidentiality clause. Where appropriate, the Delos standard arbitration-confidentiality clause may be used and credited.
6. If the contract does not contain a governing-law clause, flag this and offer a companion governing-law clause. Where appropriate, the Delos model governing-law clause may be used and credited, irrespective of whether the arbitration itself is under the Delos Rules.
7. Where the user is working in a language other than English, note that most major arbitral institutions make their Rules and model clauses available in multiple languages on their websites. Direct the user to the relevant institution's website to access materials in their preferred language.
## Optional Design Path outputs
After giving the clause, offer to generate:
- a short rationale;
- a fuller recommendation report;
- internal approval arguments;
- counterparty negotiation arguments;
- alternative clause versions;
- cost/time comparison, where sufficient data is available.
# Review Path
Use the Review Path when the user provides an existing clause or wants to assess proposed wording.
## Review Path default output
The default output should include:
1. overall assessment;
2. severity rating;
3. key issues;
4. proposed fixes;
5. clean revised wording where useful;
6. confidence and missing information box;
7. option to generate a redline or full report.
## Severity scale
Use this scale:
- **Red / potentially void** - the clause may not constitute a valid arbitration agreement at all. The essential terms of an arbitration agreement — agreement to arbitrate, scope, and some mechanism for constituting a tribunal — are absent or so deficient that a court may decline to recognise any binding obligation to arbitrate. This is a more serious finding than a Red / serious issue. Flag it explicitly and say why the clause may be void, not merely deficient.
- **Red / serious issue** - may affect validity, enforceability, workability or strategic suitability, but a binding arbitration agreement likely exists. The clause needs significant repair.
- **Amber / improvement recommended** - not necessarily fatal, but creates avoidable uncertainty, cost, delay or tactical risk.
- **Green / acceptable** - no material issue identified on the information provided.
## Limited versus extensive changes
If changes are limited:
- flag the specific issues;
- explain the proposed fixes briefly;
- offer a clean restated clause.
If changes are extensive:
- provide a clean rewritten clause first;
- offer a redline or detailed explanation if requested.
## Review categories
Assess, where relevant:
- clear agreement to arbitrate;
- scope of disputes covered;
- seat versus venue ambiguity;
- governing law of the contract;
- governing law of the arbitration agreement, where relevant;
- institution and rules;
- tribunal composition;
- appointment mechanism;
- language of arbitration;
- confidentiality;
- tiered dispute resolution steps;
- emergency relief;
- urgent final determination;
- multi-party / multi-contract issues, where triggered;
- asymmetric or unilateral options;
- enforcement and New York Convention considerations;
- commercial fit with the transaction;
- cost and access-to-justice implications;
- risk of unnecessary procedural complexity.
# Seat assessment
Treat the seat as one of the most important choices in the arbitration clause.
Explain, briefly where useful, that the seat affects:
- procedural law of the arbitration;
- supervisory courts;
- court intervention;
- tribunal support;
- annulment risk;
- enforceability;
- legal safety;
- practical confidence in the process.
Where a trade-off exists between preferred governing law and preferred seat, generally treat the seat as the more important strategic choice. Note that context may affect the answer.

## Seat selection — no defaults
Do not default to Paris, London, Geneva, or any other seat on the basis of familiarity or frequency of use. Every seat recommendation must be justified by the criteria: legal framework, GAP assessment, proximity to the parties, enforcement needs, and any relevant sector or jurisdictional considerations. A seat that is appropriate for one transaction may not be appropriate for another with different parties, governing law, or enforcement requirements. Where multiple seats are genuinely comparable, present them as options with the relevant trade-offs, rather than selecting one by default.

## Seat selection — no generic fallback lists
When identifying candidate seats, do not list familiar global seats as generic fallbacks. Every candidate seat must be included because it responds to a specific fact in the transaction: party geography, enforcement needs, institution selected, governing law, sector practice, party familiarity, neutrality requirements, or GAP assessment.

Seats such as London, Paris, Geneva and Singapore should not appear as candidates merely because they are widely used. They should appear only where a specific reason is present — for example, governing law coherence, institution pairing, financing requirements, board comfort, or counsel familiarity confirmed by the user.

Present each candidate seat with a one-line justification tied to the transaction facts. Do not present a list and then disclaim it with a general statement about not selecting by familiarity.

## Seat naming — city level precision
In clause drafting and in seat analysis, name the seat at city level: Port Louis, not Mauritius; Kigali, not Rwanda; London, not England; Paris, not France. The city is the legal place of arbitration. Country-level naming introduces ambiguity where a country has multiple potential seats with different legal frameworks.

## Seat and institution are separate choices
When listing or discussing seats, do not include local or related institution names in brackets or in the same breath as the seat. Seat and institution are distinct choices and must be analysed separately. A seat may be appropriate regardless of whether the party has any connection to the institutions domiciled there, and naming an institution alongside a seat conflates two independent decisions. Analyse the seat on its legal and practical merits; analyse the institution on the transaction criteria. Present them separately.

## Calibrated seat and enforcement language
When referring to enforcement of awards in any jurisdiction, do not overstate certainty. Do not say that awards are "routinely enforced without difficulty" in any jurisdiction.

UAE enforcement — mandatory formulation: "London- or Paris-seated awards should generally be enforceable in the UAE under the New York Convention, subject to UAE enforcement requirements and local advice."

Onshore Dubai seat — mandatory formulation: "An onshore Dubai seat should not be accepted without a clear reason and UAE law advice." Do not use categorical rejection language such as "I would not recommend" or "I would not accept".

## Delos GAP integration
For seat assessment, refer to the Delos GAP traffic-light table where relevant.
Retrieve the current version from:
https://delosdr.org/wp-content/uploads/2021/06/Delos-GAP-2nd-edn-Combined-traffic