agentsclimarketplace

Ai high stakes verifiable

Skill r-irbe/proof-skills/skills/ai-high-stakes-verifiable

APM-installable agent skills for Lean 4 and Mathlib4 — proof tactics, math domains, review and research workflows, generic tooling.

Install
npx -y skills add r-irbe/proof-skills --skill ai-high-stakes-verifiable

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

USE FOR: Formally verifiable AI, high-stakes AI systems, safety-critical deployment, certification, and regulatory compliance. Use for reasoning about AI systems that must be provably correct, auditable, or certifiable — medical AI, autonomous vehicles, legal decision support, military/intelligence systems, and mathematically rigorous AI governance. DO NOT USE FOR: formal verification proofs in Lean (use @lean-ai-formalization); agentic AI dynamics (use @ai-agentic-evolving); causal/deontic reasoning (use @ai-causal-deontic). TRIGGERS: high stakes, verifiable AI, safety critical, certification, regulatory compliance, AI assurance.

SKILL.md

4.1 KB, as published. Nobody here has run it

High-Stakes & Formally Verifiable AI

Mathematical frameworks for AI systems where correctness, safety, and auditability are non-negotiable.


Routing

  • USE FOR: Formally verifiable AI, high-stakes AI systems, safety-critical deployment, certification, and regulatory compliance. Use for reasoning about AI systems that must be provably correct, auditable, or certifiable — medical AI, autonomous vehicles, legal decision support, military/intelligence systems, and mathematically rigorous AI governance.
  • DO NOT USE FOR: formal verification proofs in Lean (use @lean-ai-formalization); agentic AI dynamics (use @ai-agentic-evolving); causal/deontic reasoning (use @ai-causal-deontic).
  • TRIGGERS: high stakes, verifiable AI, safety critical, certification, regulatory compliance, AI assurance.

Workflow

  1. Identify the verification class: certification (regulatory), assurance case (engineering), or formal property (mathematical).
  2. Pick the matching framework from the body (ISO 21448, ASIL, formal-property catalog).
  3. Produce the verification artifact appropriate for the class — assurance argument, requirement decomposition, or property statement ready for @lean-ai-formalization.
  4. Hand off: to @lean-ai-formalization for the proof, to @applied-data-information-security for security-specific properties, to @lean-zettelkasten.

Recovery & STOP

  • STOP if the property is informally argued but needs a Lean proof — hand directly to @lean-ai-formalization.
  • STOP if the system is agentic/emergent — delegate dynamics modelling to @ai-agentic-evolving first.
  • STOP if the certification target is jurisdiction-specific and not in the body — escalate to @applied-legal-reasoning.

Handoffs

  • Predecessors: agent:gateway, skill:lean-research.
  • Successors: skill:lean-ai-formalization, skill:lean-security-formalization, skill:applied-data-information-security.

Detailed reference

Full content for ai-high-stakes-verifiable lives in references/ai-high-stakes-verifiable-handbook.md. Load that file when the skill is convened; the SKILL.md only carries the dispatch contract and the parts index.

SectionTopic
Part 1Verification Hierarchy
Part 2Safety Standards & Certification
Part 3Formal Safety Analysis
Part 4Provably Robust AI
Part 5Decision Support in High-Stakes Domains
Part 6Connection to Project Lean Modules
Part 7Research Directions

See also

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.