Dense, information-rich prose that assumes competence (no explaining what a JWT, gateway, or BYOK is); every section — the surface test, naming map, marker rules — earns its place with no padding.

Gives concrete, copy-paste-ready guidance: the exact grep `grep -rn 'GRIDA-GG' editor packages desktop`, exact marker comment formats, and an identifier-by-identifier naming table — actionable for a code-organization skill.

The numbered 'Working on GG' sequence and the 'When GG crosses security' precedence (run the security skill first, then this pass) give a clear ordered process with an explicit checkpoint; no destructive batch step is present that would require a validation loop.

No bundle files exist; the body is self-contained and well-organized into clearly signaled sections with one-level-deep cross-references to companion skills (../security/SKILL.md, ../ee-billing/SKILL.md), satisfying the simple-skill allowance.

Lists multiple concrete surfaces — 'the scoped-token mint, the OpenAI-compatible + native gateway endpoints, the `gg` client provider, and the desktop wiring' — matching the multi-action anchor.

Explicitly answers what (the GG working pattern and its four surfaces) and when ('Use when adding or touching any GG file, the `gg` provider kind, the `gg:ai` token audience, or deciding whether code belongs to the gateway surface').

The 'Use when adding or touching any GG file, the `gg` provider kind, the `gg:ai` token audience' clause gives explicit triggers, but the terms are internal code identifiers/jargon rather than phrases a user would naturally say, and common variations are thin.

A clearly scoped internal niche (Grida Gateway surface) with a unique grep marker, named companions, and explicit carve-outs ('Not GG: the billing ledger...'), so it is unlikely to trigger for the wrong skill.