Would a HOL skill validation workflow be useful for whatsapp-claude-plugin?

[internet-dot]

If useful, I can open one PR that adds a validate-only HOL skill workflow.
It would only add one workflow file, keep publish mode off, and leave runtime code alone.
If that is not useful here, no problem.

[m13v]

the permission relay concept (approving/denying tool requests from WhatsApp) is clever. biggest challenge with that flow is latency. Claude's tool calls have timeouts, and if the approval comes from WhatsApp even 30 seconds late, the session might have already timed out or moved on. we handle a similar pattern by batching tool requests and presenting them as a single approval message rather than one-by-one. a validation workflow makes sense since the WhatsApp input parsing needs to be bulletproof. a typo in a tool approval could accidentally grant permissions you didn't intend. I'd keep it lightweight though, just schema checks on the message format and a confirmation echo before executing.

[m13v]

our WhatsApp MCP server that handles the bidirectional Claude/WhatsApp communication: https://github.com/m13v/whatsapp-mcp-macos

and the tool execution layer with the batched approval pattern I mentioned: https://github.com/m13v/fazm/blob/main/Desktop/Sources/Providers/ChatToolExecutor.swift