docs: add clrun install + interactive prompt handling guidance

[cybertheory]

Summary

This PR adds explicit clrun guidance for handling interactive CLI/TUI prompts in gstack skills, and does so in the template/source path required by the contributing guidelines.

What changed

• Updated shared preamble source:
  • scripts/resolvers/preamble.ts
• Added a dedicated Interactive CLI / TUI handling section that documents:
  • when to use clrun (interactive prompts, wizards, yes/no, select/checkbox UIs)
  • core execution flow (clrun <command> → clrun tail → respond)
  • text input and key navigation patterns (clrun <id> "...", clrun key ...)
  • session management (clrun status, clrun kill)
  • shell-variable quoting guidance
• Added install/fallback guidance:
  • install if missing: npm install -g clrun
  • no-global fallback: npx clrun <command>
• Regenerated generated docs from templates across hosts (SKILL.md and per-skill generated outputs).

Why

The existing docs described policy for interactive prompts but did not provide a deterministic operational workflow for terminal interactivity. This change makes agent behavior explicit and repeatable, including setup when clrun is not installed.

Validation

• Regenerated docs:
  • bun run gen:skill-docs --host all
• Verified consistency and freshness:
  • bun run skill:check
• Checks pass with expected warnings for skills that intentionally have no $B commands.