docs: add 4 resource evaluations + tasks/lessons.md pattern (v3.27.7+)

- Add tasks/lessons.md variant to Fresh Context Pattern section — lightweight
  self-improvement loop where Claude captures behavioral rules after corrections
- Resource evaluations (4 new, 67 → 78 total):
  - SKILLMIND/Boris Cherny workflow (2/5): self-improvement loop concept
  - Augment Context Engine MCP (3/5): cloud alternative to grepai
  - LinkedIn context engineering post (2/5): marginal relevance
  - LinkedIn activity-7428930570451083264 (inaccessible)
- README: update evaluations count 67 → 78, threat DB badge 18/341 → 19/655
- CLAUDE.md: add Whitepaper Generation section (PDF + EPUB commands)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

guide/ultimate-guide.md

@@ -1595,6 +1595,18 @@ while :; do cat TASK.md PROGRESS.md | claude -p ; done
 - `PROGRESS.md` — Learnings, completed tasks, blockers
 - Git commits — Each iteration commits atomically
 
+**Variant: tasks/lessons.md**
+
+A lightweight alternative for interactive sessions (no loop required): after each user correction, Claude updates `tasks/lessons.md` with the rule to avoid the same mistake. Reviewed at the start of each new session.
+
+```
+tasks/
+├── todo.md      # Current plan (checkable items)
+└── lessons.md   # Rules accumulated from corrections
+```
+
+The difference from PROGRESS.md: `lessons.md` captures *behavioral rules* ("always diff before marking done", "never mock without asking") rather than task state. It compounds over time — the mistake rate drops as the ruleset grows.
+
 | Traditional | Fresh Context |
 |-------------|---------------|
 | Accumulate in chat history | Reset per task |