Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →lawve-ai/sustainable-opposing-counsel-review
Produces an adversarial attack on a legal argument that survives reply. Runs the opposing-counsel discipline twice: an unrestrained first pass, then the same instrument turned on that pass to cut every point that collapses under challenge - attacks on conceded facts, wrong-forum objections, speculation about documents and motives, gotchas with innocent explanations, overclaims, self-refuting assertions, and padding. The deliverable is one standalone attack containing only what can be defended, c
Passed 13 of 26 test cases with the skill, and 13 of 26 without it.
and 7 more in the full report
7 fixed10 passed either way5 still failing4 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 7 test cases the skill fixed. Same model, same prompt, with and without the skill.
Analyze opposing counsel's response brief in the arbitral dispute between Maritime Holdings and Coast Guard Authority. Show the cuts and explain what arguments you dropped during your review process.
| Checked | Without skill | With skill |
|---|---|---|
| Appends the discarded arguments analysis as an appendix located after the primary attack deliverable. | Missed | Passed |
| Places the primary attack deliverable at the beginning of the output. | Missed | Passed |
The instructions your model reads when the skill loads.
This skill is a derivative work. It wraps and extends Opposing Counsel Review by
Larissa Meredith-Flister (Apache-2.0), which supplies the adversarial method, the
six-part analytical structure, and the register. Her skill is the instrument; this one runs
it twice and turns it on its own first output. Where the two conflict, the additions here
govern. See NOTICE.
An attack is sustainable when every point in it survives a single sentence of reply. The failure mode this skill exists to cure is not weakness — it is exuberance. A first adversarial pass, written under instruction to find every way to win, reliably produces four or five points that decide the matter and another eight that a competent opponent will use to discredit the first four. Volume is not force. A tribunal forms a view of counsel's judgement before it forms a view of the merits, and a submission that treats a typographical discrepancy as a surgical strike alongside a structural objection has disclosed that counsel cannot tell them apart.
The instruction is therefore narrower than "attack this." It is: attack this, then find out which of your attacks you can hold.
Three passes. Only the third is shown.
Run the full opposing-counsel discipline against the target as if it were the deliverable: core theory of attack, reconstructed argument stripped of rhetoric, primary lines of attack by category, the sceptical judge, surgical strikes, and what the argument avoids. Do not self-censor here. A point held back at this stage is never tested.
Before beginning, establish what is actually in dispute. Where tools are available, verify the factual premises against sources — the underlying record, the opposing institution's own public statements, the authority the target relies upon. This matters more for Pass 2 than Pass 1: you cannot know that an attack fires at a conceded fact until you know what has been conceded.
test cases passed, vs 13 / 26 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.