docs(competitive): cocoindex-code deep-dive (#1556)

* docs(competitive): add cocoindex-code competitive deep-dive

Semantic retrieval vs structural intelligence — different problem spaces,
complementary tools. Covers 8-principle FOUNDATION.md alignment, A–H
feature comparison, and 4 backlog candidates (doctor command, Claude Code
skill, per-language stats, embedding docs).

* docs(competitive): fix two factual inaccuracies in cocoindex-code deep-dive

- Remove duplicate "OpenCode" in choose-when list (point 7 had "Codex, OpenCode, or OpenCode")
- Fix backlog preamble section attribution: CC1/CC3 source from section G,
  not section C (Query Interface)

Author: carlos-alm