The Ledger
Luke Carriere Proposes Draft BIP to Standardize Bitcoin Co-Signer Spending Policies
Luke Carriere’s draft proposes a portable format for describing what transactions Bitcoin co-signers may approve and under what conditions. It specifies five policy components, including spending templates, evidence requirements and verification profiles, while leaving the co-signer mechanisms themselves flexible.

A draft standard for co-signer policies
Luke Carriere’s draft BIP, titled “Co-signer Spending Policies,” proposes a shared format for describing the rules that Bitcoin co-signers enforce. The document focuses on standardizing the policy, not the co-signer itself: the same format is intended to work with different ways of verifying a spend and different sources of evidence. “This document specifies a portable spending policy for outputs that include one or more policy-enforcing co-signers,” Carriere writes.
The motivation is that separate co-signer designs can each define their own policy, leaving wallets and verification tools without a common description of what a co-signer may sign. The draft points to blinded co-signers, Chain Code Delegation and predicate blind signatures as examples of approaches that use co-signers in different ways. It also notes that Liquidium and Lendasat use policy co-signers for bitcoin-backed loans, pairing multisignature arrangements with oracle-driven settlement.
Five components describe each policy
The proposal divides a policy into five parts: an output, a co-signer group, transaction templates, conditions and profiles. The output is a Taproot descriptor holding the funds. The co-signer group states which signers participate and how many must approve a spend. Templates specify the transactions that may be signed, including their exact outputs; conditions set what must be true before a template can be used. Profiles identify the verification mode, evidence formats and any accountability scheme.
The draft requires the output to use a Taproot descriptor and says each template must correspond to a script leaf requiring the co-signer group and the participants named for that template. Conditions that Bitcoin Script can enforce, such as timelocks, must be placed in the script rather than left to the co-signers. Each policy must also keep at least one spending path that does not require a co-signer.
For the co-signer group, the draft describes two options. In an aggregate group, all co-signers combine into one MuSig2 key and must sign, allowing any honest signer to block an invalid spend but making an offline signer a barrier to a valid one. In a threshold group, a specified number of keys sign through the leaf script. This can tolerate offline co-signers, but the threshold must include enough honest signers to block an invalid spend.
Templates and evidence govern what can be signed
A template must define every transaction output exactly. Its amount expressions can use literal values, named policy variables, values supplied by evidence, arithmetic operations and a remainder for the input value left after other outputs and fees. A template may have no more than one remainder output. Participants set the fee, while co-signers check each non-fee output against the template.
Conditions can combine transaction predicates, evidence conditions and time conditions. Evidence conditions specify which attestors must attest to an event or numeric value, the required threshold and the maximum age of the evidence. The draft places a limit on using co-signers for conditions Bitcoin Script can already express: those conditions belong in the script leaf.
Profiles keep verification methods and trust arrangements flexible. The specified open-v0 mode lets the co-signer see the policy, PSBT and evidence. The draft lists blinded verification and enclave-based verification as modes still to be specified. It also requires support for DLC-format oracle attestations, so existing oracles can be used without modification.
Under open-v0, a co-signer may sign only when the inputs are spent under the policy through the requested template’s leaf, the outputs match the template and its conditions are met by valid evidence within the permitted age. The co-signer must refuse otherwise and provide participants with a record of the policy ID, transaction ID, template and evidence used. The policy ID is defined as a BIP340 tagged hash of the policy’s canonical serialization.
Sources
- Original source Luke Carriere (GitHub) github.com
Related stories
- Draft Multisig Session Protocol would let remote Bitcoin cosigners set up wallets without a coordinator
- Preprint proposes journal-ordered withdrawal permits for Bitcoin custody
- Blockstream Research reviews lattice-based signatures for Bitcoin
- Researchers propose splitting Bitcoin mining rewards in fee-only era
- Delving Bitcoin user posts STARK proof for Bitcoin’s UTXO set