Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →okhp3/okhp3-equilibrium-review
Evaluate a document, report, spreadsheet, hypothesis, decision memo, or Agent Skill with independent evidence review, conditional disruption, and evidence-based adjudication.
Passed 17 of 21 test cases with the skill, and 10 of 21 without it.
and 9 more in the full report
9 fixed9 passed either way3 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
In our claim ledger for 'CLM-404', we have recorded the claim text and evidence sources. Which schema field records what operational failure occurs if the claim is false?
| Checked | Without skill | With skill |
|---|---|---|
| Identifies consequence as the schema field for failure impact. | Missed | Passed |
The instructions your model reads when the skill loads.
OverKill Hill P³ · overkillhill.com · github.com/OKHP3
The Equilibrium Review skill turns agreement and disagreement into a traceable decision. It evaluates an artifact without assuming that a polished answer, majority vote, or self-review is proof of correctness.
Use this skill as a quality gate around an artifact-producing workflow. The artifact may be a paper, research memo, spreadsheet, dashboard, Outlook assessment, hypothesis, decision recommendation, or Agent Skill. The skill's default job is to review and improve the decision record, not silently rewrite the artifact or take consequential action.
The equilibrium metaphor is operational, not a claim that language-model reviewers compute a formal Nash equilibrium. A useful review has provisional harmony when independent reviewers converge for evidence-backed reasons. It has productive disharmony when they identify a material difference. A disruptor is introduced to search for falsifiable counterexamples when harmony may be false. A negotiator resolves disagreement from evidence, not from vote counting.
| In scope | Out of scope |
|---|---|
| Claim, evidence, utility, safety, portability, and reader-usefulness review | Inventing evidence or declaring perfection |
| Conditional five-role multi-agent review | Unbounded agent debate or ceremonial dissent |
| Hypothesis and measurement-design audit | Performing a statistical test without appropriate data or expertise |
| Review records, ledgers, decision gates, and revision plans | Autonomous publication, sending messages, or changing source systems |
| Domain adapters for documents, data, reports, and skills | Replacing legal, medical, financial, security, or other specialist authority |
test cases passed, vs 10 / 21 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.