docs(claude): scope the generated-mirror rule to the files that are mirrors (#3468)

The Documentation section of `CLAUDE.md` said "never hand-edit the
generated mirrors under `docs/content/`, `skills/worktrunk/reference/`,
or `plugins/worktrunk/skills/`", which reads as "everything under these
directories is generated." That's wrong for two of the three:

- `docs/content/` — non-command docs (faq, tips-patterns, worktrunk,
claude-code, llm-commits) are hand-edited primaries; only command pages
have generated regions.
- `skills/worktrunk/reference/` — `shell-integration.md`,
`troubleshooting.md`, and `README.md` are hand-maintained sources
(already carved out in `.gitattributes` with
`linguist-generated=false`); the rest are generated mirrors.
- `plugins/worktrunk/skills/` — wholly generated, as stated.

The over-broad wording nearly steered #3455 away from editing
`shell-integration.md`, which was the correct file to edit. The reworded
sentence names the split per directory and defers to `docs/CLAUDE.md`'s
sync taxonomy (already linked at the end of the paragraph) for the
per-file answer. No behavior or generated-file changes.

> _This was written by Claude Code on behalf of Maximilian Roos_

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

Author: max-sixty

CLAUDE.md

@@ -43,7 +43,7 @@ Load relevant skills before starting; reload when scope changes mid-session. Pro
 
 ## Documentation
 
-Behavior changes require doc updates. `src/cli/mod.rs` (`after_long_help` plus clap attributes) is the PRIMARY SOURCE for command pages; never hand-edit the generated mirrors under `docs/content/`, `skills/worktrunk/reference/`, or `plugins/worktrunk/skills/`. Ask: "does `--help` still describe what the code does?" `cargo test --test integration test_docs_are_in_sync` checks doc sync; editing help text (`after_long_help`, `about`, arg docs) also changes the rendered `--help` snapshots, which that test leaves untouched — `cargo insta test --accept -- --test integration "test_help"` regenerates them (the pre-merge hook runs both). Sync taxonomy, help-text authoring (three render contexts, link text, config-TOML blocks): `docs/CLAUDE.md`.
+Behavior changes require doc updates. `src/cli/mod.rs` (`after_long_help` plus clap attributes) is the PRIMARY SOURCE for command pages; their rendered mirrors in `docs/content/` and `skills/worktrunk/reference/` are generated, as is all of `plugins/worktrunk/skills/` — but both directories also hold hand-edited primaries (non-command docs in `docs/content/`, skill-only pages like `shell-integration.md` in the reference dir), so check which file is primary in the sync taxonomy before editing. Ask: "does `--help` still describe what the code does?" `cargo test --test integration test_docs_are_in_sync` checks doc sync; editing help text (`after_long_help`, `about`, arg docs) also changes the rendered `--help` snapshots, which that test leaves untouched — `cargo insta test --accept -- --test integration "test_help"` regenerates them (the pre-merge hook runs both). Sync taxonomy, help-text authoring (three render contexts, link text, config-TOML blocks): `docs/CLAUDE.md`.
 
 ## Plugin Layout