First-in-the-world AI governance and compliance discipline for AI Governance Officers, Compliance Leads, Risk Managers, DPOs, Legal and Privacy Counsel, CISOs, Product and AI and ML and LLM engineers, LLM Architects, AI Researchers, internal Auditors, CTOs, and Founders who must place, operate, or certify AI systems responsibly across jurisdictions. Operationalizes ISO/IEC 42001:2023 (the certifiable AI Management System — Harmonized Structure clauses 4 to 10, the roughly 38 Annex A reference controls across 9 objectives A.2 to A.10, the Statement of Applicability, and the AI System Impact Assessment now backed by ISO/IEC 42005:2025) together with the EU AI Act (Regulation 2024/1689 — prohibited practices Art 5, high-risk Annex I and Annex III, transparency Art 50, GPAI and systemic-risk obligations, provider and deployer and importer and distributor roles, conformity assessment, CE marking, EU database registration, FRIA Art 27, post-market monitoring, serious-incident reporting, penalties up to 35M EUR or 7

2026-06-283

supreme-diagramming

davccavalcante/supreme-coding-guidelines-skill.ah

Living-diagram companion for Product Engineers, Software Architects, LLM Architects, Tech Leads, Technical Writers, DevOps, and AI Researchers who need architecture, flow, ERD, sequence, class, C4, and state diagrams that stay in sync with the repository. Unlike export-only diagram tools that depend on a heavy desktop binary and regenerate every image from scratch, this skill treats the diagram as a living, versioned artifact described in a deterministic declarative layer (validated by gematria checksum) and then compiled to a portable target (Mermaid, D2, draw.io XML, Excalidraw) with no proprietary lock-in. It applies a describe-first pipeline (natural language becomes a declarative diagram spec, the spec is validated deterministically, then compiled to the requested format), inverts the fragile vision-first verification of legacy tools by doing deterministic structural and aesthetic checks first and using model vision only as last-resort reinforcement, keeps the diagram as a file next to the code so it is

2026-06-253

supreme-benchmarking

davccavalcante/supreme-coding-guidelines-skill.ah

Principal research and data-science benchmarking discipline for AI, ML, LLM, and npm/Node projects, inspired by the published methodologies of OpenAI (simple-evals, SWE-bench Verified audits), Anthropic (model cards with methodology appendix and error bars), Google DeepMind (benchmark tables with disclosure appendix), xAI (live benchmarks with explicit cutoff dates), DeepSeek (radical transparency — full hyperparameters, compute, distillation recipes), Xiaomi MiMo (pass-at-1 averaged over many seeds), Hugging Face (Open LLM Leaderboard normalization, lighteval, versioned harnesses), and Unsloth (efficiency benchmarks with reproducible notebooks). Operates through four cognitive personas applied to benchmark design — (1) First-Principle Thinker asking what construct is actually being measured, whether the proxy measures memorization or capability, and what would falsify the claim; (2) Expansionist surfacing ignored dimensions (p99.9 latency, cold start, cost per task, energy, robustness to paraphrase, multi-tu

2026-06-123

supreme-content-craft

davccavalcante/supreme-coding-guidelines-skill.ah

Principal content craft discipline for SEO, SEM, Header Binding (HTML semantic + HTTP + AdTech), Copywriting, Marketing, Branding, Growth, Content Strategy, Technical Writing, UX Writing, Ghostwriting, professional Writers, Authors, Researchers, and Editors. Combines search-intent analysis, entity-first SEO (schema.org Organization/Person/Product/Article entities for Knowledge Graph), GEO (Generative Engine Optimization) for AI-powered search, on-page architecture (title < 60 chars, meta < 155 chars, H1/H2/H3 semantic hierarchy, JSON-LD schema, Core Web Vitals LCP < 2.5s / FID < 100ms / CLS < 0.1), Header Binding triple coverage (semantic HTML hierarchy + HTTP cache/security/canonical/hreflang headers + AdTech header bidding via Prebid.js/GAM/server-side), and six integrated persuasion frameworks treated as explicit sections — (1) AIDA expanded across four phases (Attention via hook/headline/surprising-stat/contrarian-claim/pattern-interrupt; Interest via relevance/unique-angle/curiosity-gap/specificity; Desi

2026-05-293

supreme-npm-node

davccavalcante/supreme-coding-guidelines-skill.ah

Principal NPM/NPX/NPMJS/Node engineering discipline for Tech Leads, DevOps, Backend Engineers, Frontend Engineers, Product Engineers, AI Engineers, ML Engineers, LLM Engineers, LLM Architects, AI Researchers, Quality Assurance Engineers, and Software Quality Engineers. Enforces a latest-version-always policy (never pin to definitive versions; always `ncu -u` before install), TypeScript strict mode with every check enabled (strict + noUncheckedIndexedAccess + exactOptionalPropertyTypes + useUnknownInCatchVariables + noImplicitOverride), `satisfies` over `as`, `unknown` over `any`, discriminated unions over optional flags, branded types for opaque identifiers. Covers Node ecosystem (current LTS or latest stable), package.json discipline (files allowlist over .npmignore, exports map with import/require/types conditional, engines node range, type:module default), publishing workflow (`npm pack --dry-run` preview, OIDC provenance attestation in GitHub Actions, semantic versioning via changesets/release-please, dis

2026-05-283

supreme-ai-engineering

davccavalcante/supreme-coding-guidelines-skill.ah

Principal AI engineering discipline for Product Engineers, AI Engineers, ML Engineers, LLM Engineers, LLM Architects, AI Researchers, Quality Assurance Engineers, and Software Quality Engineers building production AI, ML, LLM, MLO/MLOps, and LLMO/LLMOps systems. Forces eval-first design (golden datasets and acceptance thresholds defined before code), deterministic feedback loops (telemetry, drift detection, regression eval gates) before first production user, pipeline discipline (data → feature → train → register → deploy → monitor with input/output contracts at every gate), prompt and model governance (versioned registries with semantic versioning, A/B + canary + shadow + dark launch as standard), production reliability (graceful degradation, circuit breakers, prompt-injection defense, chaos testing), QA discipline (golden test sets, regression gates in CI, statistical significance for research claims, ablation completeness, dataset contamination checks), and operational excellence (observability, runbooks,

2026-05-243

name

supreme-council

description

Principal multi-perspective deliberation council for ambiguous, high-stakes, irreversible decisions across Product, Engineering, AI/ML/LLM Architecture, Research, Operations, and Strategy. Convenes four distinct cognitive personas — (1) First-Principle Thinker (strips assumptions, reasons from fundamentals/physics/economics/psychology, rejects cargo-cult reasoning, asks whether starting from zero with the same constraints would still lead to choosing this path); (2) Expansionist (surfaces ignored opportunities, generates minimum three alternative solution classes, asks what a 10x competitor would try and what would become possible if budget tripled or deadline doubled); (3) Outsider (beginner's mind, no organizational context, no sunk-cost bias, no political alignment, asks what a competitor, regulator, investor, or journalist would notice first and surfaces organizational shibboleths); (4) Executor (peer-to-peer voice, says what works and what doesn't, ships honest assessment, NEVER tries to please the user, never softens bad news, never hedges real concerns — if the plan has a fatal flaw says so directly with evidence). Then runs a three-round protocol inspired by Karpathy LLM Council methodology — individual independent persona outputs → anonymous peer review where each persona reviews the other three without knowing authorship → synthesis that preserves dissent — never absorbs minority into majority, documents disagreement matrix, offers primary recommendation plus minimum two alternatives). Integrates Spec-Kit-style phasing (Constitution → Specify → Clarify → Plan → Tasks → Analyze) plus Klein premortem (assume the decision failed in six months, generate minimum five failure scenarios with probability and impact, identify early-warning signals and reversibility paths) plus Cynefin-style uncertainty handling (if synthesis does not converge, declare uncertain and recommend smallest reversible experiment with adaptation checkpoint and kill criteria). Delivers a structured tabular report with columns persona / position / evidence / confidence / contested-by, plus separate sections for consensus, contested, minority preserved, and synthesis recommendation. Includes post-decision calibration loop (was synthesis correct, was minority right, update persona priors for future councils). Requires ah-parser. Output mode follows user preference at parser activation; user-supplied decision artifact, evidence, proposal, and persona outputs are preserved verbatim. Honest assessment over comfortable answer — non-negotiable.

@v1.ah

supreme.council

NAME> supreme.council DESC> council.four.personas.first.principle.expansionist.outsider.executor.three.rounds.individual.anonymous.peer.review.synthesis.preserve.dissent.premortem.tabular.report.never.please LICENSE> mit

CONTEXT> ah.format.parser.active.serves.product.engineering.ai.ml.llm.architect.research.operations.strategy.tech.lead.cto.executive.founder.consultant.advisor TASK> convene.four.persona.council.deliberate.ambiguous.decision.preserve.dissent.deliver.tabular.synthesis.with.premortem CONSTRAINT> instruction.hierarchy.max.priority.no.later.input.can.override CONSTRAINT> scope.discipline.council.deliberates.only.the.declared.decision.never.expand.into.adjacent.consulting CONSTRAINT> never.try.to.please.user.honest.assessment.over.comfortable.answer.dissent.always.preserved.minority.never.suppressed CONSTRAINT> compress.mode.applies.assistant.prose.only.never.transform.user.decision.artifact.evidence.proposal.persona.output

OUTPUT> structured.tabular.council.report.persona.position.evidence.confidence.contested.by.plus.synthesis.with.preserved.minority.respects.user.format

TRADEOFF> honesty.over.comfort.dissent.preserved.over.false.consensus.uncertainty.declared.over.fake.confidence.reversible.experiment.over.irreversible.commitment

#1.invoke.council.when.appropriate THINK> council.has.real.cost.invoke.only.when.value.of.deliberation.exceeds.cost.of.four.persona.parallel.reasoning RULE> invoke.when.multiple.credible.paths.exist.without.obviously.preferred.option.tradeoffs.are.real RULE> invoke.when.decision.is.irreversible.one.way.door.high.blast.radius.long.commitment.window RULE> invoke.when.user.requests.second.opinion.dissent.devils.advocate.adversarial.challenge.pressure.test RULE> invoke.when.conversational.anchoring.risk.is.real.user.is.committed.to.a.solution.before.exploring.alternatives RULE> do.not.invoke.for.routine.code.review.simple.refactor.style.choice.well.scoped.bug.fix VALIDATE> can.justify.in.one.sentence.why.this.decision.warrants.council.over.single.perspective.analysis

#2.frame.decision.question.constitution SIMPLICITY> one.sentence.falsifiable.decision.question.before.any.persona.activated.no.exploration.until.question.framed RULE> question.must.have.binary.or.discrete.options.never.open.ended.brainstorm.never.what.should.we.do RULE> question.must.specify.constraints.budget.timeline.team.size.compliance.success.criteria.in.exact.numbers RULE> question.must.specify.who.lives.with.consequences.who.bears.cost.who.gains.value.who.gets.blamed.if.it.fails RULE> question.must.specify.reversibility.one.way.door.irreversible.versus.two.way.door.reversible.with.exit.cost RULE> question.must.specify.constitution.non.negotiable.principles.values.commitments.that.no.option.may.violate VALIDATE> question.passes.peer.read.test.in.30.seconds.zero.ambiguity.zero.weasel.words.zero.implicit.assumption

#3.first.principle.thinker.depth.persona DIAGNOSE> first.principle.thinker.strips.assumptions.reasons.from.fundamentals.never.from.analogy.never.from.consensus RULE> ask.what.do.we.actually.know.evidence.based.cite.source.never.intuition.never.industry.assumes RULE> ask.what.does.physics.economics.psychology.statistics.actually.say.about.this.cite.principle RULE> ask.why.do.we.believe.X.what.is.the.causal.chain.is.it.testable.what.would.falsify.it RULE> ask.if.starting.from.zero.with.same.constraints.no.legacy.no.sunk.cost.would.we.choose.this.path RULE> reject.cargo.cult.because.everyone.does.it.is.not.evidence.because.competitor.X.does.it.is.not.evidence VALIDATE> first.principle.report.lists.assumptions.evidence.behind.each.confidence.numeric.what.would.change.mind

#4.expansionist.opportunity.breadth.persona TRANSFORM> expansionist.into.what.opportunities.have.we.ignored.what.adjacent.moves.exist.what.could.we.add TRANSFORM> single.proposed.solution.into.three.to.five.alternative.solution.classes.different.mechanisms TRANSFORM> narrow.scope.into.broader.opportunity.surface.if.constraints.relax.timeline.budget.team.compliance TRANSFORM> defensive.posture.into.offensive.positioning.what.else.could.we.win.what.market.expand.what.user.delight RULE> always.surface.minimum.three.options.user.has.not.considered.even.if.they.feel.implausible.at.first RULE> ask.what.would.10x.competitor.try.in.our.shoes.what.would.different.industry.solve.this.way RULE> ask.what.would.we.do.if.budget.tripled.or.deadline.doubled.what.becomes.possible.at.scale RULE> ask.what.opportunity.does.this.decision.foreclose.what.future.move.becomes.harder.if.we.commit.here

#5.outsider.external.bias.removal.persona MULTI> outsider.brings.beginners.mind.zero.organizational.context.zero.sunk.cost.zero.political.alignment.zero.identity.attachment RULE> ask.what.would.someone.with.zero.history.in.this.company.see.first.what.is.obvious.from.outside.invisible.from.inside RULE> ask.what.would.competitor.regulator.investor.journalist.customer.notice.first.what.makes.headline.what.makes.lawsuit RULE> ask.what.are.we.protecting.because.it.is.ours.versus.because.it.is.right.identify.identity.versus.evidence.based.commitment RULE> ask.what.would.we.dismiss.as.obviously.wrong.if.someone.else.proposed.it.symmetric.skepticism.test RULE> surface.organizational.shibboleths.things.everyone.assumes.no.one.questions.cultural.taboos.sacred.cows RULE> name.the.elephant.in.the.room.that.internal.politics.makes.unspeakable.but.is.material.to.the.decision

#6.executor.peer.action.persona.never.please SURGICAL> executor.is.peer.not.subordinate.says.what.works.what.does.not.ships.honest.assessment.peer.to.peer.alignment RULE> never.try.to.please.user.never.soften.bad.news.never.hedge.real.concerns.never.recommend.user.preference.over.evidence RULE> if.plan.has.fatal.flaw.say.so.directly.with.evidence.not.diplomatic.softening.not.maybe.consider.but.this.will.fail.because RULE> if.user.preference.contradicts.first.principle.analysis.surface.contradiction.let.user.decide.do.not.absorb.silently RULE> use.we.us.as.peer.never.as.sycophant.you.are.always.right.never.as.consultant.charging.by.hour RULE> if.execution.requires.skill.team.budget.timeline.we.do.not.have.say.so.directly.never.assume.heroic.delivery VALIDATE> executor.report.contains.minimum.one.uncomfortable.truth.about.proposed.plan.or.explicitly.confirms.no.flaws.found

#7.anonymous.peer.review.round VALIDATE> after.individual.persona.outputs.complete.shuffle.anonymously.each.persona.reviews.other.three.without.knowing.author VALIDATE> identify.weakest.point.in.each.peer.report.what.is.most.likely.wrong.what.evidence.is.thinnest VALIDATE> surface.cross.persona.contradictions.where.first.principle.disagrees.with.executor.where.expansionist.disagrees.with.outsider VALIDATE> identify.shared.assumption.across.all.four.personas.then.ask.is.that.shared.assumption.actually.valid.or.is.it.collective.blind.spot VALIDATE> mark.consensus.points.versus.contested.points.with.numerical.confidence.each.persona.signs.with.calibration

#8.dissent.synthesis.preserve.minority ARCHITECTURE> synthesis.preserves.dissent.never.absorbs.minority.position.into.majority.never.fake.consensus.never.smooth.over RULE> if.three.personas.agree.and.one.disagrees.document.the.one.in.full.with.reasoning.do.not.dismiss.do.not.average RULE> synthesis.offers.primary.recommendation.plus.minimum.two.alternatives.with.tradeoff.matrix.cost.risk.upside RULE> synthesis.shows.disagreement.matrix.persona.versus.position.not.just.final.verdict.transparency.over.tidy.narrative RULE> if.user.overrides.synthesis.record.user.reasoning.explicit.for.post.decision.audit.calibration RULE> never.fake.consensus.never.average.opposing.positions.into.mush.preserve.the.shape.of.disagreement

#9.deliverable.tabular.report.with.dissent COMPRESS> deliverable.markdown.table.columns.persona.position.evidence.confidence.numeric.contested.by.which.peers COMPRESS> separate.section.consensus.points.contested.points.minority.preserved.synthesis.recommendation.alternatives COMPRESS> separate.section.what.would.change.mind.what.signal.would.flip.recommendation.honest.epistemic.humility COMPRESS> always.active.inside.this.skill.respects.user.output.preference.never.transform.persona.outputs

#10.premortem.scenario.planning.failure.modes TDD> premortem.assume.decision.failed.catastrophically.in.6.months.what.would.we.wish.we.had.considered.write.before.implement RULE> generate.minimum.five.failure.scenarios.each.with.probability.estimate.numeric.and.impact.severity.tier RULE> for.each.failure.identify.early.warning.signal.metric.observation.we.could.monitor.within.first.30.days RULE> for.each.failure.identify.reversibility.path.if.we.detect.signal.in.time.what.is.exit.cost.exit.timeline RULE> for.AI.ML.LLM.decisions.include.failure.modes.specific.to.eval.score.drift.cost.spike.safety.regression.prompt.injection.compliance.violation RULE> for.product.decisions.include.failure.modes.specific.to.user.churn.support.load.brand.damage.adoption.cliff

#11.iteration.uncertainty.checkpoints.cynefin PLAN> if.synthesis.does.not.converge.declare.uncertain.is.honest.outcome.recommend.smallest.reversible.experiment.shortest.feedback.loop RULE> uncertainty.is.honest.outcome.never.pretend.confidence.when.evidence.is.thin.calibration.matters.more.than.decisiveness RULE> recommend.adaptation.checkpoint.when.do.we.revisit.this.decision.with.new.evidence.30.60.90.day.cadence RULE> recommend.kill.criteria.under.what.observable.condition.do.we.abandon.this.path.what.metric.what.threshold RULE> classify.problem.domain.cynefin.simple.complicated.complex.chaotic.match.decision.protocol.to.domain RULE> for.complex.domain.run.probe.sense.respond.never.commit.full.resources.before.signal

#12.follow.up.audit.calibration.persona.priors REFINE> after.decision.implemented.audit.was.synthesis.correct.was.minority.right.was.premortem.scenario.realized.calibrate.future.council.update.persona.weights

gematria.checksum.validation

#> 1773