docs: add Xquik context and fix pipe-mode status

[kriptoburak]

Summary

• document Xquik REST API and MCP output as optional X post metadata context
• keep watch-cli authoritative for video frames, audio, and transcripts
• defer media dependency checks until pipe mode receives a valid URL
• preserve failed work-item exit codes in the aggregate pipe result
• make the manual Claude skill copy command portable across operating systems

Independent watch-cli Fixes

watch --pipe checked yt-dlp and other media dependencies before reading stdin. Empty streams and malformed URLs therefore returned a missing-dependency error even though neither operation needs media tooling. Dependency checks now run lazily on the first valid URL, preserving deterministic empty-input and validation behavior.

The pipe loop also used if ! process_url; then RC=$?, which captured the negated condition's success status instead of the failed command's status. Failed downloads could therefore leave the aggregate exit code at 0. The loop now captures process_url directly and preserves its real status.

The README copied a symlinked skill directory. GNU cp -r preserves that relative symlink, which becomes broken after installation. It now copies the canonical root SKILL.md directly.

Validation

• bash -n across all binaries, installer, and schema test
• warning-level ShellCheck on changed shell files
• bash tests/test-output-schema.sh, including a new deterministic missing-yt-dlp pipe test
• all command help smoke checks
• MCP server npm ci and TypeScript build
• README and SKILL link checks
• custom portable-skill frontmatter and marketplace path checks
• git diff --check
• Xquik API and MCP documentation HTTP checks
• full audit of comments, reviews, review threads, checks, commits, conflicts, upstream, and open issues

Xquik is an independent third-party service. Not affiliated with X Corp. "Twitter" and "X" are trademarks of X Corp.