Skip to content

Support MCP caching metadata (ttlMs/cacheScope) in vMCP #5761

Description

@JAORMX

Part of #5743 (Workstream C: 2026-07-28 stateless readiness).

Work

2026-07-28 requires caching metadata on list/read results: ttlMs + cacheScope (public/private) via CacheableResult (SEP-2549), and tools/list SHOULD be deterministically ordered for cache hits.

  • vMCP: emit ttlMs/cacheScope on aggregated list results — natural fit, the caching aggregator already has a TTL; ensure deterministic tool ordering.
  • vMCP client: respect backend-provided TTLs in the aggregation cache instead of a fixed TTL.
  • Proxies: pass through untouched (verify nothing strips _meta/result fields).

Depends on: the go-sdk v1.7 adoption issue for types. Priority: opportunistic — pairs well with the vMCP stateless work.

Metadata

Metadata

Assignees

No one assigned

    Labels

    needs-triageIssue needs initial triage by a maintainer

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions