chrisdoc/hevy-mcp hevy-mcp@6.1.4

on GitHub

Patch Changes

• #1036 a9a7594 Thanks @chrisdoc! - Clarify canonical MCP tool names and require tool parameters to be sent in the arguments object.

• #1033 331a3bc Thanks @chrisdoc! - Increase the default Hevy API operation deadline to accommodate slow, large collection responses.

• #1033 331a3bc Thanks @chrisdoc! - Return a structured, confirmed acknowledgement from create-routine, including the authoritative routine when Hevy provides one.

• #1035 b17ab5b Thanks @chrisdoc! - Guard MCP telemetry session ID generation when node:crypto.randomUUID is unavailable.

• #1032 ad391b1 Thanks @chrisdoc! - Treat Hevy's HTTP 400 workout validation responses as actionable input errors instead of generic API failures.

• #1031 ae58ddd Thanks @chrisdoc! - Use endpoint-agnostic guidance for HTTP 409 conflicts so routine conflicts are not described as body measurement conflicts.

• #1036 a9a7594 Thanks @chrisdoc! - Improve the error guidance when updating a routine that no longer exists in Hevy.

• #1034 1a6b016 Thanks @chrisdoc! - Improve the startup guidance when non-loopback HTTP mode is missing its bearer token.

• #1030 1bae2ba Thanks @chrisdoc! - docs: clarify Hevy API authentication failures

• #1045 cd46318 Thanks @chrisdoc! - Reduce per-request CPU in the Worker: memoize tool schema registration once per isolate, and preload the actual compact-JSON-Schema conversions at module scope so the first request only reads already-converted schemas. This eliminates the repeated conversion that caused intermittent Worker exceeded CPU time limit (503) responses.

• #1042 31bf76a Thanks @chrisdoc! - Adapt telemetry and Worker tracing fallbacks to the upgraded Cloudflare Workers types and Node type definitions.