[michalmaj]

Summary

First docs-only polish batch from the final polish/UX audit. Exactly the 4 scoped fixes — no functional changes, no code/fixture/config touched.

1. README.md — L1 mission count was stale ("20 focused exercises", actually 23). Opening sentence ("each with focused missions and two checkpoint projects before the boss fight") incorrectly generalized L2-L7's P01/P02+boss-fight pattern onto Level 1, which actually has 5 independent boss fights. Reworded to a true general statement; all other level counts (L2-L7: 10/7/6/9/8/8 missions) were already correct and left untouched.
2. COURSE_MAP.md — Level 4's prerequisite said only "Level 3 complete," omitting the mandatory Data Recall Checkpoint that sits between them. Now reads "Level 3 complete, Data Recall Checkpoint complete." No other prerequisite lines touched; existing plain (non-blockquote) style preserved.
3. level_5_maintainability/missions/04_pytest_fixtures/README.md — the only checker-invocation command in the entire course missing uv run (every sibling L5 mission uses uv run python check.py). Fixed both occurrences (the Step 3 instruction and the matching checklist line) for internal consistency. Verified empirically: uv run python check.py runs correctly from the mission directory (fails only on the expected incomplete-scaffold state, not on command/path).
4. level_7_concurrency_and_background_work/projects/03_concurrent_tournament_runner/README.md — the L7 boss fight (end of all 7 levels) had no pointer to the mandatory Final Analytics Epilogue, the one case in the whole course where this matters most. Added a 2-sentence bridge; link target verified to exist.

Also checked: the audit's reported vscode-webview:// artifact in L7 links — confirmed via grep -rn across all tracked Markdown that no such literal URL exists anywhere in the repo; it was a rendering/audit artifact, not a real issue. Nothing to fix.

Explicitly out of scope (future batches)

L3 starter intro, L5/L6 boss-fight summaries, prerequisite formatting unification, Next: link-style sweep, Key Insight cleanup, AI-smell edits, em-dash cleanup, top-level README rewrite.

Test plan

• All 4 changed files are .md — zero .py/.json/.csv/config changes (git diff --name-only)
• L1 counts in README now match filesystem (23 missions / 5 projects)
• uv run python check.py verified to run correctly from level_5_maintainability/missions/04_pytest_fixtures/
• L7→Epilogue link target (final_analytics_epilogue/README.md) verified to resolve
• tools/author_check.py → 750 checks passed, 0 errors (unchanged from before)
• tools/course_status.py → output unchanged
• No .progress/__pycache__/temp artifacts left in the working tree

🤖 Generated with Claude Code