LegalOSS133.3ktracked
Submit a project
Agent Skills & Plugins

pre-motionMaintained

b1rdmania/pre-motion

Adversarial premortem for England & Wales civil litigation — a Claude Skill. Builds the strongest version of a case, then attacks it from four angles to find where it loses.

GitHub
Agent Skills & Plugins

pre-motion

Adversarial premortem for England & Wales civil litigation. Builds the strongest version of a case, then attacks it from four angles — procedural, substantive, evidentiary, strategic — to find where it loses before opposing counsel does.

For solicitors stress-testing before issue, in-house counsel before sign-off, mediators valuing settlement, and litigation funders pricing a matter.

Listed on lawve.ai (Litigation & Dispute Resolution).

Install

git clone https://github.com/b1rdmania/pre-motion ~/.claude/skills/pre-motion

Or in a Legalise workspace: add it from the skill library — review the manifest, grant capabilities, enable on a matter, run from chat. Every run leaves a signed, auditable record.

Usage

/pre-motion
/pre-motion --depth=fast
/pre-motion --depth=thorough

Run it against a matter with the facts, evidence references, claim heads, and the strongest version of the case as you see it. It returns a ranked stress-test brief.

What it does

  • Builds the optimistic baseline — the strongest version the evidence supports.
  • Inspects the evidence: document gaps, cross-document contradictions, timeline holes, each flagged with a severity.
  • Runs four adversarial passes (parallel sub-agents where supported, otherwise sequential), each told the case has been lost and asked to walk back why — one per failure category.
  • Synthesises a brief: ranked failure scenarios with severity, likelihood and mitigation; evidence inconsistencies; blind spots; settlement-posture implications; and one one-sentence verdict.
  • Marks every uncertain point inline — [CITE NEEDED], [SME VERIFY], [EVIDENCE FLAG] — so nothing reads as settled.

What it doesn't do

  • Predict the outcome — it surfaces failure modes, not results.
  • Decide whether to take, settle, or withdraw — those are counsel's calls.
  • Replace a formal counsel opinion. It is a structured prompt for that conversation, not a substitute.
  • Cover non-UK procedure (US federal, Scotland, Northern Ireland).
  • Verify case law against a live source — check any authority it cites before relying on it.

Requirements

  • Claude Code or Claude Cowork. No MCP connectors required.
  • A matter to run against (facts and evidence references). In a host workspace, the CPR 31.22 / privilege gate is enforced upstream; run standalone only on material you are permitted to use.

License

Apache-2.0.