agentleFS
mobeetlemobeetle/legalsuite/skills/legal-dispute/SKILL.md

legal-dispute

Skill1 starsChanged 27 days agoApache-2.0View on GitHub

Disputes, litigation and arbitration for any legal system, claimant or defendant side: pre-action demands, claims received, court and arbitral proceedings, appeals, and internal investigations. Builds the sourced chronology and the claims-and-defences matrix (elements, burden, standard, facts, evidence, status), maps the procedure with every deadline and preclusion point, assesses merits, quantum and costs with likelihood bands, sets strategy (litigate, settle, ADR, wait), plans submissions, prepares hearings and witnesses, and runs document review with privilege logs. Use for a dispute, lawsuit, claim, defence, demand letter, payment order, appeal, arbitration, settlement, evidence, witness, hearing or investigation (CS: spor, žaloba, předžalobní výzva, vyjádření k žalobě, platební rozkaz, odpor, odvolání, dovolání, rozhodčí řízení, důkazy, svědek, jednání, smír, narovnání, interní šetření). Load alongside the legal engine.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 5 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read ../legal/SKILL.md first if the legal engine isn't loaded.

    line 24
  2. Very specific:

    No profile (or a placeholder one): offer setup, or run provisional — mark it [PROVISIONAL — configure for tailored output] …

    line 27
  3. Distinctive:

    Deadlines first, always: Every procedural deadline is computed conservatively and shown with its rule, trigger and basis, never as a bare …

    line 31
  4. Distinctive:

    Settlement communications: follow the jurisdiction's rules (pack §8 (Civil procedure & relief requirements), pack §11 (Evidence, burden of …

    line 42
  5. Very specific:

    Every element shows its burden and its evidence.

    line 60
  6. Very specific:

    Consent: Ask, never infer a load-bearing gap.

    line 63
  7. Distinctive:

    every factual statement mapped to a fact row and its evidence designation.

    line 114
  8. Distinctive:

    every argument mapped to a matrix element.

    line 115
  9. Distinctive:

    Start here when something arrived: a demand letter, a claim, a payment order, an arbitration request, a decision.

    line 129

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

legal-dispute

Object of work. A claim, defence, pre-action demand, proceeding or internal investigation.

Imports. Obey the capability files named in the front-matter metadata.imports; do not restate them. This file carries only the mode's delta.

The Disputes hat. Litigation differs from transactional work in exactly the ways a document map doesn't capture: an adversary acting in real time, hard deadlines and preclusion, burden and standard of proof, evidence rules, costs risk, and strategy that turns on facts more than on drafting. This skill runs on the fact map (../legal/references/fact-map.md).

Read ../legal/SKILL.md first if the legal engine isn't loaded. Procedure, deadlines, costs and evidence rules come from the jurisdiction pack: pack §6 (Limitation & time bars), pack §7 (Computing time & delivery rules), pack §8 (Civil procedure & relief requirements), pack §9 (Costs & fees), and pack §11 (Evidence, burden of proof & privilege).

No profile (or a placeholder one): offer setup, or run provisional — mark it [PROVISIONAL — configure for tailored output] (../legal/references/capabilities/configure.md).

Guardrails specific to disputes

  • Deadlines first, always. Every procedural deadline is computed conservatively and shown with its rule, trigger and basis, never as a bare date. A received document starts a clock: find it before anything else.
  • Nothing is filed, sent or served by the suite. Each one is an approval.
  • Burden and proof are surfaced, not decided. Whether evidence meets the standard is the lawyer's judgment. Mark elements disputed or missing instead of proven on thin evidence.
  • Witness statements in the witness's own words. Outlines and questions only. Never script testimony or "improve" a recollection.
  • Authenticity. Flag evidence whose authenticity is open (unsigned copies, screenshots, audio/video that could be manipulated, altered metadata).
  • Settlement communications follow the jurisdiction's rules (pack §8 (Civil procedure & relief requirements), pack §11 (Evidence, burden of proof & privilege)). Mark them, keep them out of submissions unless the rules allow, and don't assume foreign doctrines (e.g. US FRE 408) apply.

AST obligation — non-negotiable

Deadlines & Case Intake (Phase 1) and Chronology (chronology.md, Phase 2) feed the Claims & Defences Matrix (claims-matrix.md, Phase 3) and the Dispute AST. Once the chronology and claims are framed, destructure the case into the AST: Party/Event/Object are the facts, Norm the legal basis, Control Flow the elements of each claim, and a Trace per defence or counterclaim (theirs and ours). Print the full tree in one fenced text box; every claim, defence and assessment cites its node. Edges carry the evidence and burden links.

Grammar and structure: ../legal/references/ir-and-provenance.md.

Urgency. Deadlines first, always — a missed preclusion is the whole case. Every element shows its burden and its evidence; mark disputed/missing, not proven, on thin proof. Never cite authority you did not retrieve.

Consent. Ask, never infer a load-bearing gap; a write outside the matter folder needs an approved manifest (gate.md).

Phases (cluster via the Gate Plan; enter at any phase)

  1. Deadlines & case intake. From legal-intake (qualify gate, then its capture, parties and urgency steps) or run now. Side (claimant / defendant / both). Forum options: court or arbitration (clause?), international jurisdiction and venue (pack §8 (Civil procedure & relief requirements); EU pack for cross-border), applicable law (the pack's private-international-law rules), and enforceability of a win where the assets are. Time bars (pack §6 (Limitation & time bars)) for every claim, our own and the other side's possible counterclaims.
  2. Chronology (chronology.md, a view of facts.md). Events with date precision, source, who asserts them, contested or not. After the table: inconsistencies (sources that disagree), gaps (periods or events with no document), missing documents, and critical documents (the 5–10 that decide the case).
  3. Claims and defences matrix & Dispute AST (claims-matrix.md, a view of the Issues table). Fed by the Phase 1 intake and Phase 2 chronology: for each claim and each defence (ours and theirs), map the legal basis [AUTH:A…], its elements, the burden for each element (who must assert and prove it, pack §11 (Evidence, burden of proof & privilege)), the standard, the facts per element, the evidence per fact, the status, and the Dispute AST. This is the core artifact. Every later phase reads it.
  4. Procedural map (procedural-map.md). Forum and stage; every deadline, with its computation rule and basis (pack §7 (Computing time & delivery rules), to deadlines.md); preclusion points (the last moment to plead facts and offer evidence, limits on new facts in appeal, pack §8 (Civil procedure & relief requirements)); fees and costs regime (pack §9 (Costs & fees)); interim measures available and their conditions; the path to enforcement.
  5. Merits, quantum and risk (case-assessment.md). Per claim: likelihood band and its drivers, amount range (best/likely/worst), costs exposure (own, adverse, court fees per pack §9 (Costs & fees)), duration, and enforceability. Include a structured Quantum, Interest & Costs Exposure schedule:

    | Head of claim | Principal | Interest / penalty rate & trigger | Accrued to date | Court / arbitral fees | Adverse costs exposure | Net expected value (likelihood × quantum − costs) |

    Then run the challenge capability (../legal/references/capabilities/challenge.md) with the judge and opposing counsel lenses. Its verdicts go into the assessment (challenge.md, alias oponentura.md), and are not summarised away.

  6. Strategy. Options: litigate, settle, ADR (mediation, arbitration if available), do nothing or wait. For each: BATNA, costs, time, risk, reputational and relationship effects. Then a decision tree with the 2–3 decisive uncertainties, and a settlement range (walk-away and target) for the lawyer to set. Then the first moves: e.g. a pre-action demand (pack §8 (Civil procedure & relief requirements): required content and waiting period for costs), preservation, interim relief, or a set-off declaration. Decisions go to decisions.md.
  7. Submissions. Claim, defence, reply, appeal, extraordinary remedies, arbitration request, demand letter, settlement agreement: drafted by the draft mode (../legal-review/references/modes/draft.md, contentious architecture) against this door's plan:
    • the required particulars and the exact relief sought (exact prayer for relief / order sought, e.g. petit) per pack §8 (Civil procedure & relief requirements), worded so it's enforceable if granted;
    • every factual statement mapped to a fact row and its evidence designation;
    • every argument mapped to a matrix element;
    • the other side's arguments answered element by element. Then trace.py on the annotated version, the verify capability (../legal/references/capabilities/verify.md) before filing, and scan.py outbound on the clean version.
  8. Hearing and witnesses. references/hearing-prep.md: hearing plan, witness outlines tied to the theory, exhibit list, anticipated objections, cross-examination points from the inconsistencies.
  9. Document review (investigations, arbitration document production, a court order to produce, foreign proceedings): references/document-review.md covers responsiveness and privilege coding, the privilege log, and the production set.

Received claim or demand (defence side)

Start here when something arrived: a demand letter, a claim, a payment order, an arbitration request, a decision.

  1. Identify the document, how and when it was delivered (statutory e-delivery / service fictions (e.g. registered post, court portal, data box), pack §7 (Computing time & delivery rules)), and the deadline it triggers with its rule. That becomes the first line of the first card.
  2. What exactly is claimed, on what basis, with what evidence. Check it against the time bars (pack §6 (Limitation & time bars): is the claim already time-barred, and must that be raised as an objection?).
  3. The immediate options and their deadlines: respond, contest (e.g. opposition to a payment order), pay or settle, counterclaim or set-off.
  4. Then continue with phases 2–6.

Internal investigations

Same fact map, a different object: allegations instead of claims. Lifecycle and safeguards (scope, independence, interviews, employee rights, whistleblower protection, data protection, privilege, reporting duties): references/investigations.md.

Outputs

chronology.md · claims-matrix.md · procedural-map.md · case-assessment.md (with challenge.md, alias oponentura.md) · strategy.md · submissions in drafts/ · exhibits.md · privilege-log.md (if document review) · and deadlines.md and decisions.md kept current throughout.

Checks

  • trace on every submission, assessment and chronology: facts resolve and have evidence; contested facts are visible.
  • grounding on every authority, ours and the other side's.
  • Deadlines: each has a rule, a trigger, a basis and an owner.
  • the verify capability (../legal/references/capabilities/verify.md) before anything is filed or sent; the challenge capability (../legal/references/capabilities/challenge.md) on every case assessment.

Fan-out

Fan out the fact work: one read-only packet per claim or element, and per witness. Authority retrieval runs in its own wave. The parent owns the claims matrix, the strategy and every submission. Protocol: ../legal/references/capabilities/orchestrate.md.

What this mode does not do

  • Predict the outcome as a certainty: it gives a likelihood band with its drivers, never a guarantee.
  • File, send or decide to litigate or settle — those are the client's approvals.
  • Build the case before the procedure: the procedural map and deadlines come first.

Counter-example (illustration)

Illustration, not law. Told "we'll win", the memo adopts that framing. Mistake to catch: a confident verdict replaced the claims-and-defences matrix and the burden of proof. Correct: state the elements, who must prove each, the evidence, and a likelihood band — not an assurance.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.