A Formally Verified Static Analysis Framework for Compositional Contracts

Abstract

A commercial or financial contract is a mutual agreement to exchange resources such as money, goods and services amongst multiple parties. It expresses which actions may, must and must not be performed by its parties at which time, location and under which other conditions. We present a general framework for statically analyzing digital contracts, formal specifications of contracts, expressed in Contract Specification Language (CSL). Semantically, a CSL contract classifies traces of events into compliant (complete and successful) and noncompliant (incomplete or manifestly breached) ones. Our analysis framework is based on compositional abstract interpretation, which soundly approximates the set of traces a contract denotes by an abstract value in a lattice. The framework is parameterized by a lattice and an interpretation of contract primitives and combinators, satisfying certain requirements. It treats recursion by unrestricted unfolding. Employing Schmidt's natural semantics approach, we interpret our inference system coinductively to account for infinite derivation trees, and prove their abstract interpretation sound. Finally, we show some example applications: participation analysis (who is possibly involved in a transfer to whom; who does definitely participate in a contract) and fairness analysis (bounds on how much is gained by each participant under any compliant execution of the contract). The semantics of CSL, the abstract interpretation framework and its correctness theorem, and the example analyses as instances of the abstract interpretation framework have all been mechanized in the Coq proof assistant.

Extracting smart contracts tested and verified in Coq

“…Apart from the example in Figure 3, we successfully applied the developed extraction to several variants of the counter contract, to the crowdfunding contract described in [ANS20] and to an interpreter for a simple expression language. The latter example shows the possibility of extracting certified interpreters for domain-specific languages such as Marlowe [LST18], CSL [HLM20] and the CL language [BBE15,AE18]. This represents an important step towards safe smart contract programming.…”

Extracting functional programs from Coq, in Coq

“…4, we successfully applied the developed extraction to several variants of the counter contract, to the crowdfunding contract described in (Annenkov et al, 2020), the contracts from Sections 6 and 7 and to an interpreter for a simple expression language. The latter example shows the possibility of extracting certified interpreters for domain-specific languages such as Marlowe (Lamela Seijas and Thompson, 2018), CSL (Henglein et al, 2020) and the CL language (Bahr et al, 2015;Annenkov and Elsman, 2018). This represents an important step towards safe smart contract programming.…”

“…However, for certain domain-specific contract languages (e.g. Marlowe (Lamela Seijas and Thompson, 2018), CSL (Henglein et al, 2020), CL (Bahr et al, 2015;Annenkov and Elsman, 2018)) the approach of (Šinkarovs and Cockx, 2021) looks promising. It would be interesting to reproduce the approach in Coq, with the additional benefit of reasoning about the semantics preservation using the MetaCoq formalisation.…”

Extracting functional programs from Coq, in Coq

“…The latter example shows the possibility of extracting certified interpreters for domain-specific languages such as Marlowe (Lamela Seijas & Thompson 2018), CSL (Henglein et al . 2020) and the CL language (Bahr et al . 2015; Annenkov & Elsman 2018).…”

“…Marlowe (Lamela Seijas & Thompson 2018), CSL (Henglein et al . 2020), CL (Bahr et al . 2015; Annenkov & Elsman 2018), the approach of Šinkarovs & Cockx (2021) looks promising.…”