proof-writer

To write rigorous mathematical proofs from rough theorem sketches, you provide the theorem statement, assumptions, and a proof sketch to normalize the claim, triage feasibility, and output a structured proof document with explicit dependency mapping and step-by-step justification.

Proof feasibility triage evaluates whether a claim is provable as stated, requires weakening or adding assumptions, or remains unjustified. For overly strong claims, it outputs a corrected claim accompanied by a valid proof instead of fabricating missing steps.

Yes, you can formalize machine learning theory lemmas requiring boundary-case reasoning and careful quantifier handling. The skill produces rigorous arguments for propositions and corollaries about ML/AI claims by tracking dependencies across mathematical steps.

Structuring a complete proof package involves normalizing the theorem statement, performing feasibility triage, and writing a document containing explicit assumptions, notation, strategy, a dependency map, and step-by-step justification to a target file.

When your theorem statement is too strong to prove as written, the skill repairs the claim instead of fabricating justification. It outputs a corrected, weaker claim accompanied by a rigorous proof that can be mathematically justified.

Yes, you need YAML frontmatter metadata for discovery and processing when generating mathematical proofs for ML theory claims. The skill requires this metadata alongside your theorem statements and assumptions to produce a rigorous proof package.