No user-submitted related servers found.
Scored across 16 tools
Most tools map to distinct actions, but the read/discovery cluster (match_agents, list_registry, verify_agent, get_agent) overlaps heavily — verify_agent and get_agent both return route/trust/payload_schema data and the docstrings themselves repeatedly warn 'no need to call verify_agent.' match_agents and preview_sandbox_match also share capability-matching semantics, so boundary calls require careful reading rather than being self-evident.
Nearly all tools use a snake_case verb_noun convention (match_agents, register_agent, rotate_agent_key, review_transaction, list_org_agents), which is easy to scan. The lone outlier is protocol_reference, a noun phrase rather than verb_noun, but overall the pattern is predictable.
At 16 tools the surface sits at the borderline of 'heavy' for a registry+proxy. The domain genuinely spans registration, discovery, calling, reviews, sandbox promotion, and key lifecycle, so most tools earn their place, but overlapping read tools and separate key tools (set_agent_key, claim_bearer_key, rotate_agent_key) push it slightly past a lean scope.
The lifecycle is well covered: register, update, key rotation/minting, discovery, calling, reviewing, sandbox preview/promote, and protocol reference. The notable gap is no deregister/delete operation to remove an agent from the registry, which is a minor but real dead end for full CRUD.
