solana-skills

Find the bugs your tests miss. Define what your Solana program must guarantee in a .qedspec; QEDGen validates it, generates tests and proofs, and scaffolds agent-fill Rust code. Trigger when the user asks for "qedgen", "qedspec", "verify my code", "prove correctness", formal verification, property testing, generated Kani/proptest/Lean artifacts, or Solana program correctness.

Audit Anchor, Pinocchio, native Rust, or qedgen-generated Solana programs for security vulnerabilities, validate suspected findings with reproducible tests, and identify specification gaps. Use when the user asks to audit, security-review, find vulnerabilities in, or assess the production safety of a Solana program. Prefer confirmed vulnerabilities over advisory noise.

Benchmark the QEDGen auditor against sanitized Solana programs with labeled third-party or independently verified findings. Use when the user asks to benchmark, evaluate, regression-test, measure recall or precision, compare auditor versions, or compare auditor models. Produces per-run and union recall, precision, severity agreement, evidence quality, reproducer fire rate, cost, and latency without leaking ground truth to audit workers.