@circom:tree.proof

Generate the Poseidon Merkle inclusion proof for the leaf at the given index, as a [pathIndex siblings] pair ready for destructuring, or [pathIndex siblings length] with pad:<n>, which zero-pads lean siblings to the fixed length circuits expect. Fixed-depth proofs always have exactly depth siblings; lean proofs skip levels without one and compress the path index accordingly.

⚗️ Experimental — available at next.evmcrispr.com.

Returns: array

Syntax

@circom:tree.proof(leaves index lean:<value> depth:<value> pad:<value>)

Arguments

Name	Type	Description
leaves	array	Array of field-element leaves, in insertion order
index	number	Zero-based position of the leaf to prove
lean:	bool	lean:true — Semaphore v4 LeanIMT (the default when depth: is not set)
depth:	number	depth:<n> — zero-padded fixed-depth tree
pad:	number	pad:<n> — zero-pad lean siblings to a fixed length and append the real proof length

Examples

# Prove membership: destructure the proof and verify it against the root
set $leaves [1234 5678 9012]
set $root @circom:tree.root($leaves)
set [$index $siblings] @circom:tree.proof($leaves 1)
print "Valid:" @circom:tree.verify($root 5678 $index $siblings)

Notes

• The result is a [pathIndex siblings] pair — destructure it with set [$index $siblings] @circom:tree.proof(...).
• For fixed-depth trees and complete lean trees the path index equals the leaf index. When a lean level has no sibling, that level is skipped in both siblings and the packed path index (zk-kit generateProof().index semantics) — always thread the returned path index into circuits and @circom:tree.verify instead of reusing the leaf position.
• Lean circuits usually take a fixed-size sibling array: pad $siblings with zeros up to the circuit depth and pass the sibling count separately, as the circuit expects.

See Also

• @circom:tree.root — the tree conventions
• @circom:tree.verify — verify the generated proof