[sappenin]

This PR generates the HostBindings trait found in host_bindings_trait.rs from the definition in the rippled repo. Because the version found in rippled is not quite suitable for usage by the stdlib, we must generate a version from the authoritative source (as opposed to using that source directly).

This PR also removes the "host audit" scripts and tooling because those are no longer needed (instead of checking for conformance, we generate instead).

Summary

• tools/generate-host-bindings parses rippled's crates/xrpl-host-functions/src/lib.rs at the commit pinned in XRPLD_DOCKER_IMAGE and generates host_bindings_trait.rs (rippled docs verbatim) + host_bindings_list.rs (for_each_host_function!).
• The wasm / empty / test implementations are now macro_rules! callbacks over that list and no longer change when a host function is added.
• scripts/generate-host-bindings.sh --check is the CI drift gate (run-all.sh + new check_host_bindings_generated job); the regex audit (host-function-audit.sh, compareHostFunctions.js) is removed.
• every_host_function_has_a_default_expectation (macro-expanded from the generated list) fails cargo test if a listed host function has no default mock expectation.

This PR Supersedes #293.

Behaviour

Import names and wire signatures are unchanged (63/63 match today's hand-written set). Only trait parameter names changed (derived from rippled: out_buff_ptr→out_ptr, cache_num→cache_idx, …); all call sites are positional. A module-level #[allow(clippy::too_many_arguments)] on host_bindings_trait in host/mod.rs replaces the old per-method allows.

Follow-ups

• Migrate generate-sfields.sh / generate-tx-flags.sh / generate-ledger-objects.sh from branch HEAD to the same commit pin.
• e2e probes for the 10 host functions no contract imports today.

Test plan

• CI green including check_host_bindings_generated

[ripple-code-reviewer]

This diff only touches CI workflow config, workspace docs, and the root Cargo.toml exclude list (the actual generator code and generated files aren't in this diff chunk). The test.yml change correctly swaps the disabled regex-based host_function_audit job for an active check_host_bindings_generated job, and AGENTS.md/Cargo.toml updates are consistent with the stated design. No correctness, security, or infrastructure issues found in the shown hunks; can't verify claims about scripts/tools not included in this diff (e.g., clippy.sh/fmt.sh coverage of the new tooling workspace) so those aren't flagged per prior guidance against asserting on unseen source.

[ripple-code-reviewer]

This diff only touches xrpl-common-stdlib/src/host/mod.rs, and the changes are limited to: (1) rewording/expanding the module-level doc comment to describe the new codegen pipeline, (2) adding #[allow(clippy::too_many_arguments)] to the host_bindings_trait module declaration, and (3) adding an unconditional include!("host_bindings_list.rs") before the three cfg-gated implementation includes, with updated inline cfg comments. The ordering is correct (macro_rules! must be defined before use, and it's placed before the three includes that consume it), the cfg gates are unchanged in logic (just reworded comments), and the lint allow is scoped appropriately to replace the old per-method allows as described in the PR. I don't see any correctness, security, or maintainability issues introduced by this small, mechanical change.

[ripple-code-reviewer]

This diff only touches xrpl-common-stdlib/src/host/mod.rs, updating documentation comments to reflect the new codegen pipeline, replacing per-method clippy allows with a single module-level #[allow(clippy::too_many_arguments)], and inserting an early include!("host_bindings_list.rs") so the for_each_host_function! macro is defined before the three cfg-gated includes that consume it. The change is low-risk: it's comment rewording plus a straightforward reordering of includes with a documented rationale (macro_rules! visibility requires definition-before-use). No logic changes are visible in this diff, and the actual generated files (host_bindings_trait.rs, host_bindings_list.rs) and the generator tool itself are not part of this diff, so their correctness can't be assessed here.