Skill gap: JSON parsing/serialization in Motoko

Gap

We identified a gap in the current skill coverage: no skill covers JSON parsing or
serialization in Motoko. mo:core has no JSON support, so any task that reads or
produces JSON (HTTP request bodies, external API payloads, config) requires a third-party
mops package, and no skill points at a recommended approach. Agents currently discover one
by luck (mops search) or hand-roll parsing.

Context

Parsing/serializing JSON in Motoko today means pulling in a mops package such as
edjcase/json. Which package to recommend, and the basic parse/serialize pattern, is the
kind of thing a skill should settle so agents don't guess.

For discussion — where should this live?

• A section or pitfall in the motoko skill recommending a mops JSON package + a short usage pattern?
• Or a note in mops-cli?
• Probably does not warrant a standalone skill.

Note: the motoko skill tracks the caffeinelabs upstream, so a motoko-skill change here
may need to be routed upstream rather than kept as an icskills-owned section — worth
deciding as part of placement.

Filing for discussion on placement before anyone writes content.

Start by reviewing the existing motoko and mops-cli skills, then compare the caffeinelabs upstream placement and the edjcase/json package. Resolve whether the guidance belongs in motoko or mops-cli, including upstream ownership, before defining the recommended package and usage content.