Add Google Gemini exports as a second generated platform

Proves the PLATFORMS registry extends cleanly: adds Gemini (Gem instructions)
alongside ChatGPT, generated from the same SKILL.md source.

- scripts/build-exports.mjs: register `gemini` -> exports/gemini/<bundle>/<skill>/
  GEM_INSTRUCTIONS.md (body + a one-line role primer from the description).
- Fix: the root exports/README.md now always lists every registered platform,
  so `--platform x` no longer drops the others from the overview.
- exports/gemini/: 172 generated Gem instruction files + index.
- README "Ready-to-use exports" and CHANGELOG now list Gemini.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016JWn5jRD5tcEFKrubjQ6Px

CHANGELOG.md

@@ -12,9 +12,10 @@ each new wave of skills bumps the **major** version, extensions and fixes bump
 ### Added
 - **Multi-platform export generator.** `scripts/build-exports.mjs` renders every skill
   into platform-ready files under `exports/` from a single source of truth (the
-  `SKILL.md` body), so content is never maintained twice. Ships ChatGPT
-  (`exports/chatgpt/<bundle>/<skill>/SYSTEM_PROMPT.md`) and a `PLATFORMS` registry that
-  makes adding Gemini/Cursor/etc. a few lines. Includes a `--check` mode and a
+  `SKILL.md` body), so content is never maintained twice. Ships **ChatGPT**
+  (`exports/chatgpt/.../SYSTEM_PROMPT.md`) and **Google Gemini**
+  (`exports/gemini/.../GEM_INSTRUCTIONS.md`) exports, plus a `PLATFORMS` registry that
+  makes adding Cursor/etc. a few lines. Includes a `--check` mode and a
   `check-generated` CI workflow that fails if exports or `web/skills.json` drift.
 - **Programmatic helpers (stdlib Python) for three flagship skills.** Each runs with
   zero dependencies and computes part of the work instead of estimating by hand: