fix(running-example): consistent lab-script paths, tasks.json policy, no command collisions

- Lab-script path convention: course scripts live in modules/NN/lab/; copy the
  one a step needs into tasks-app, then run by bare name (M4/M6/M7/M26 + headers).
- tasks.json stays gitignored: M20 verifies via `cli.py list`/`cat tasks.json`
  (not git diff) and frames runtime state as deliberately ignored; M22 cleanup
  uses `rm tasks.json`; M10 review-lab gets its own .gitignore.
  Module 21's lab deliberately ships NO .gitignore (teaching device) — untouched.
- Stop running-example command collisions: M5 clear->search, M6 count/clear->
  stats/purge, M7 clear/count->wipe/remaining (README + scripts + agent prompts +
  branch/worktree names). M6 conflict still reproduces on the carried usage line.

Closes #7
Closes #10
Closes #11

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TfzV5QvtPDz8LJS3Pu5VLT

modules/04-getting-the-ai-out-of-the-browser/lab/verify.sh

@@ -7,9 +7,11 @@
 # and the other two remain. This is a behavior check on the multi-file change — it does not
 # care HOW the AI implemented it, only that `delete` works end to end.
 #
-# Run it from inside your tasks-app project directory:
-#     bash lab/verify.sh
+# Copy this into your tasks-app project directory, then run it from there:
+#     cp /path/to/modules/04-getting-the-ai-out-of-the-browser/lab/verify.sh .
+#     bash verify.sh
 #
+# (It self-locates cli.py, so it also still works if you run it in place as `bash lab/verify.sh`.)
 # It saves and restores your real tasks.json, so your actual task list is left untouched.
 
 set -euo pipefail