Bitcoin researchers are advancing covenant implementation without touching the network's consensus rules. A May 2024 paper by Jeremy Rubin, titled "FE'd Up Covenants," introduces functional encryption (FE) to embed spending constraints directly into a transaction's signing key.
Rubin, the original author of the OP_CTV proposal (BIP-119), developed this approach to avoid the complex coordination required for Bitcoin soft forks. Changing Bitcoin's consensus rules demands agreement across node operators, miners, developers and economic actors—a process that has historically stalled.
Under the FE-based covenant model, a private key's signing capability is restricted to a predefined set of valid transactions. This differs from ordinary public-key cryptography, where a private key can sign any message. Bitcoin nodes verify these signatures using the existing OP_CHECKSIG operation, requiring no protocol-level change.
Functional encryption extends public-key encryption beyond full decryption. A master secret issues subsidiary functional keys, each designed to reveal only a specific function of plaintext, not the plaintext itself. A holder of a functional key for function `f` and ciphertext of secret `x` can compute `f(x)` but cannot reveal `x` as plaintext. This enforces specific conditions without exposing the underlying secret.
Academically, FE was formalized in 2011 by Dan Boneh, Amit Sahai, and Brent Waters. Constructions for arbitrary functions followed in 2013, though early versions carried polynomial runtime and computational overheads far exceeding current technology, relying on new, untested multilinear maps. Progress accelerated in 2021 when researchers developed FE constructions based on well-established cryptographic assumptions, opening the door to practical implementations.
In Rubin's Bitcoin covenant design, the function `f` acts as a signing function, producing a Bitcoin signature for a specific transaction. The signature depends on both a private key and the transaction itself, necessitating at least two inputs. For covenants permitting a single spending transaction (similar to OP_CTV), two inputs suffice. More expressive covenant constructions require additional arguments, such as Rubin's `f(k, tx, program, proof)` model.
Critically, the function `f` does not directly sign with the supplied private key `k`. Instead, it derives a new key pair by tweaking `k` with a hash of the covenant's conditions. This cryptographically enforces spending constraints before any signature is generated, maintaining Bitcoin's security model while adding programmable transaction logic.

